ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 807 - (show annotations)
Sun Dec 18 10:03:38 2011 UTC (9 years, 9 months ago) by ph10
File size: 223513 byte(s)
Error occurred while calculating annotation data.
Renamed isnumber in pcre_compile to avoid a clash with ctype.h in Macs, and 
fixed a bug in fixed-length calculation for lookbehinds that would show up only 
in quite long subpatterns.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.22
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 the code for calculating the fixed length of lookbehind
11 assertions.
14 Version 8.21 12-Dec-2011
15 ------------------------
17 1. Updating the JIT compiler.
19 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
20 are added as well.
22 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
23 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
24 calling _pcre_jit_exec. Some extra comments are added.
26 4. (*MARK) settings inside atomic groups that do not contain any capturing
27 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
28 was introduced by change 18 for 8.20.
30 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
31 ECMA-262 standard.
33 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
34 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
35 This bug was probably introduced by change 9 of 8.13.
37 7. While fixing 6 above, I noticed that a number of other items were being
38 incorrectly rejected as "not fixed length". This arose partly because newer
39 opcodes had not been added to the fixed-length checking code. I have (a)
40 corrected the bug and added tests for these items, and (b) arranged for an
41 error to occur if an unknown opcode is encountered while checking for fixed
42 length instead of just assuming "not fixed length". The items that were
43 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
44 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
45 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
47 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
48 being incorrectly compiled and would have given unpredicatble results.
50 9. A possessively repeated subpattern with minimum repeat count greater than
51 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
52 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
53 the first (A) could occur when it should not.
55 10. Add a cast and remove a redundant test from the code.
57 11. JIT should use pcre_malloc/pcre_free for allocation.
59 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
60 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
61 is anything other than /usr, -L is still shown).
63 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
65 14. Perl does not support \N without a following name in a [] class; PCRE now
66 also gives an error.
68 15. If a forward reference was repeated with an upper limit of around 2000,
69 it caused the error "internal error: overran compiling workspace". The
70 maximum number of forward references (including repeats) was limited by the
71 internal workspace, and dependent on the LINK_SIZE. The code has been
72 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
73 the default depends on LINK_SIZE. There is a new upper limit (for safety)
74 of around 200,000 forward references. While doing this, I also speeded up
75 the filling in of repeated forward references.
77 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
78 incorrectly expecting the subject to contain another "a" after the start.
80 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
81 in the match, the SKIP should be ignored. This was not happening; instead
82 the SKIP was being treated as NOMATCH. For patterns such as
83 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
84 tested.
86 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
87 now much more compatible with Perl, in particular in cases where the result
88 is a non-match for a non-anchored pattern. For example, if
89 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
90 "m", where previously it did not return a name. A side effect of this
91 change is that for partial matches, the last encountered mark name is
92 returned, as for non matches. A number of tests that were previously not
93 Perl-compatible have been moved into the Perl-compatible test files. The
94 refactoring has had the pleasing side effect of removing one argument from
95 the match() function, thus reducing its stack requirements.
97 19. If the /S+ option was used in pcretest to study a pattern using JIT,
98 subsequent uses of /S (without +) incorrectly behaved like /S+.
100 21. Retrieve executable code size support for the JIT compiler and fixing
101 some warnings.
103 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
104 not work when the shorter character appeared right at the end of the
105 subject string.
107 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
108 systems.
110 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
111 output it when the /M option is used in pcretest.
113 25. The CheckMan script was not being included in the distribution. Also, added
114 an explicit "perl" to run Perl scripts from the PrepareRelease script
115 because this is reportedly needed in Windows.
117 26. If study data was being save in a file and studying had not found a set of
118 "starts with" bytes for the pattern, the data written to the file (though
119 never used) was taken from uninitialized memory and so caused valgrind to
120 complain.
122 27. Updated RunTest.bat as provided by Sheri Pierce.
124 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
126 29. Computation of memory usage for the table of capturing group names was
127 giving an unnecessarily large value.
130 Version 8.20 21-Oct-2011
131 ------------------------
133 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
134 a POSIX class. After further experiments with Perl, which convinced me that
135 Perl has bugs and confusions, a closing square bracket is no longer allowed
136 in a POSIX name. This bug also affected patterns with classes that started
137 with full stops.
139 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
140 captured substring, but while checking the failing first alternative,
141 substring 1 is temporarily captured. If the output vector supplied to
142 pcre_exec() was not big enough for this capture, the yield of the function
143 was still zero ("insufficient space for captured substrings"). This cannot
144 be totally fixed without adding another stack variable, which seems a lot
145 of expense for a edge case. However, I have improved the situation in cases
146 such as /(a)(b)x|abc/ matched against "abc", where the return code
147 indicates that fewer than the maximum number of slots in the ovector have
148 been set.
150 3. Related to (2) above: when there are more back references in a pattern than
151 slots in the output vector, pcre_exec() uses temporary memory during
152 matching, and copies in the captures as far as possible afterwards. It was
153 using the entire output vector, but this conflicts with the specification
154 that only 2/3 is used for passing back captured substrings. Now it uses
155 only the first 2/3, for compatibility. This is, of course, another edge
156 case.
158 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
159 main code base, and can be used by building with --enable-jit. When this is
160 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
161 runtime --no-jit option is given.
163 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
164 ovector, the return from the function was zero, implying that there were
165 other matches that did not fit. The correct "exactly full" value is now
166 returned.
168 6. If a subpattern that was called recursively or as a subroutine contained
169 (*PRUNE) or any other control that caused it to give a non-standard return,
170 invalid errors such as "Error -26 (nested recursion at the same subject
171 position)" or even infinite loops could occur.
173 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
174 computing the minimum length on reaching *ACCEPT, and so ended up with the
175 wrong value of 1 rather than 0. Further investigation indicates that
176 computing a minimum subject length in the presence of *ACCEPT is difficult
177 (think back references, subroutine calls), and so I have changed the code
178 so that no minimum is registered for a pattern that contains *ACCEPT.
180 8. If (*THEN) was present in the first (true) branch of a conditional group,
181 it was not handled as intended. [But see 16 below.]
183 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
184 Sheri Pierce.
186 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
187 the first byte in a match must be "a".
189 11. Change 17 for 8.13 increased the recursion depth for patterns like
190 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
191 pattern contains any instances of (*THEN). If it does not, the old
192 optimizations are restored. It would be nice to do this on a per-group
193 basis, but at the moment that is not feasible.
195 12. In some environments, the output of pcretest -C is CRLF terminated. This
196 broke RunTest's code that checks for the link size. A single white space
197 character after the value is now allowed for.
199 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
200 For "fr", it uses the Windows-specific input and output files.
202 14. If (*THEN) appeared in a group that was called recursively or as a
203 subroutine, it did not work as intended. [But see next item.]
205 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
206 pattern fragments (but not containing any | characters). If A and B are
207 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
208 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
209 D. In other words, Perl considers parentheses that do not contain any |
210 characters to be part of a surrounding alternative, whereas PCRE was
211 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
212 differently. PCRE now behaves in the same way as Perl, except in the case
213 of subroutine/recursion calls such as (?1) which have in any case always
214 been different (but PCRE had them first :-).
216 16. Related to 15 above: Perl does not treat the | in a conditional group as
217 creating alternatives. Such a group is treated in the same way as an
218 ordinary group without any | characters when processing (*THEN). PCRE has
219 been changed to match Perl's behaviour.
221 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
222 RunGrepTest script failed.
224 18. Change 22 for version 13 caused atomic groups to use more stack. This is
225 inevitable for groups that contain captures, but it can lead to a lot of
226 stack use in large patterns. The old behaviour has been restored for atomic
227 groups that do not contain any capturing parentheses.
229 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
230 suppress the check for a minimum subject length at run time. (If it was
231 given to pcre_exec() or pcre_dfa_exec() it did work.)
233 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
234 fail to work when decoding hex characters in data strings in EBCDIC
235 environments.
237 21. It appears that in at least one Mac OS environment, the isxdigit() function
238 is implemented as a macro that evaluates to its argument more than once,
239 contravening the C 90 Standard (I haven't checked a later standard). There
240 was an instance in pcretest which caused it to go wrong when processing
241 \x{...} escapes in subject strings. The has been rewritten to avoid using
242 things like p++ in the argument of isxdigit().
245 Version 8.13 16-Aug-2011
246 ------------------------
248 1. The Unicode data tables have been updated to Unicode 6.0.0.
250 2. Two minor typos in pcre_internal.h have been fixed.
252 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
253 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
254 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
256 4. There were a number of related bugs in the code for matching backrefences
257 caselessly in UTF-8 mode when codes for the characters concerned were
258 different numbers of bytes. For example, U+023A and U+2C65 are an upper
259 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
260 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
261 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
262 2-byte code at the end of the subject (it thought there wasn't enough data
263 left).
265 5. Comprehensive information about what went wrong is now returned by
266 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
267 as the output vector has at least 2 elements. The offset of the start of
268 the failing character and a reason code are placed in the vector.
270 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
271 now returned is for the first byte of the failing character, instead of the
272 last byte inspected. This is an incompatible change, but I hope it is small
273 enough not to be a problem. It makes the returned offset consistent with
274 pcre_exec() and pcre_dfa_exec().
276 7. pcretest now gives a text phrase as well as the error number when
277 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
278 failure, the offset and reason code are output.
280 8. When \R was used with a maximizing quantifier it failed to skip backwards
281 over a \r\n pair if the subsequent match failed. Instead, it just skipped
282 back over a single character (\n). This seems wrong (because it treated the
283 two characters as a single entity when going forwards), conflicts with the
284 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
285 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
286 has been changed.
288 9. Some internal refactoring has changed the processing so that the handling
289 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
290 time (the PCRE_DOTALL option was changed this way some time ago: version
291 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
292 which was always a bit of a fudge. It also means that there is one less
293 argument for the match() function, which reduces its stack requirements
294 slightly. This change also fixes an incompatibility with Perl: the pattern
295 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
297 10. More internal refactoring has drastically reduced the number of recursive
298 calls to match() for possessively repeated groups such as (abc)++ when
299 using pcre_exec().
301 11. While implementing 10, a number of bugs in the handling of groups were
302 discovered and fixed:
304 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
305 (a|)*(?1) gave a compile-time internal error.
306 ((a|)+)+ did not notice that the outer group could match an empty string.
307 (^a|^)+ was not marked as anchored.
308 (.*a|.*)+ was not marked as matching at start or after a newline.
310 12. Yet more internal refactoring has removed another argument from the match()
311 function. Special calls to this function are now indicated by setting a
312 value in a variable in the "match data" data block.
314 13. Be more explicit in pcre_study() instead of relying on "default" for
315 opcodes that mean there is no starting character; this means that when new
316 ones are added and accidentally left out of pcre_study(), testing should
317 pick them up.
319 14. The -s option of pcretest has been documented for ages as being an old
320 synonym of -m (show memory usage). I have changed it to mean "force study
321 for every regex", that is, assume /S for every regex. This is similar to -i
322 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
323 using it. It makes it easier to run collections of tests with and without
324 study enabled, and thereby test pcre_study() more easily. All the standard
325 tests are now run with and without -s (but some patterns can be marked as
326 "never study" - see 20 below).
328 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
329 restoration of the capturing data to the outer values was not happening
330 correctly.
332 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
333 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
334 pattern had matched an empty string, and so incorrectly returned a no
335 match.
337 17. There was optimizing code for the last branch of non-capturing parentheses,
338 and also for the obeyed branch of a conditional subexpression, which used
339 tail recursion to cut down on stack usage. Unfortunately, now that there is
340 the possibility of (*THEN) occurring in these branches, tail recursion is
341 no longer possible because the return has to be checked for (*THEN). These
342 two optimizations have therefore been removed. [But see 8.20/11 above.]
344 18. If a pattern containing \R was studied, it was assumed that \R always
345 matched two bytes, thus causing the minimum subject length to be
346 incorrectly computed because \R can also match just one byte.
348 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
349 was incorrectly computed.
351 20. If /S is present twice on a test pattern in pcretest input, it now
352 *disables* studying, thereby overriding the use of -s on the command line
353 (see 14 above). This is necessary for one or two tests to keep the output
354 identical in both cases.
356 21. When (*ACCEPT) was used in an assertion that matched an empty string and
357 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
359 22. When an atomic group that contained a capturing parenthesis was
360 successfully matched, but the branch in which it appeared failed, the
361 capturing was not being forgotten if a higher numbered group was later
362 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
363 group 1 to "a", when in fact it should be unset. This applied to multi-
364 branched capturing and non-capturing groups, repeated or not, and also to
365 positive assertions (capturing in negative assertions does not happen
366 in PCRE) and also to nested atomic groups.
368 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
369 subject after a captured substring, to make it easier to tell which of a
370 number of identical substrings has been captured.
372 24. The way atomic groups are processed by pcre_exec() has been changed so that
373 if they are repeated, backtracking one repetition now resets captured
374 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
375 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
376 "aaa". Previously, it would have been "a". As part of this code
377 refactoring, the way recursive calls are handled has also been changed.
379 25. If an assertion condition captured any substrings, they were not passed
380 back unless some other capturing happened later. For example, if
381 (?(?=(a))a) was matched against "a", no capturing was returned.
383 26. When studying a pattern that contained subroutine calls or assertions,
384 the code for finding the minimum length of a possible match was handling
385 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
386 group 1 called group 2 while simultaneously a separate group 2 called group
387 1). A stack overflow occurred in this case. I have fixed this by limiting
388 the recursion depth to 10.
390 27. Updated RunTest.bat in the distribution to the version supplied by Tom
391 Fortmann. This supports explicit test numbers on the command line, and has
392 argument validation and error reporting.
394 28. An instance of \X with an unlimited repeat could fail if at any point the
395 first character it looked at was a mark character.
397 29. Some minor code refactoring concerning Unicode properties and scripts
398 should reduce the stack requirement of match() slightly.
400 30. Added the '=' option to pcretest to check the setting of unused capturing
401 slots at the end of the pattern, which are documented as being -1, but are
402 not included in the return count.
404 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
405 compiled something random. Now it gives a compile-time error (as does
406 Perl).
408 32. A *MARK encountered during the processing of a positive assertion is now
409 recorded and passed back (compatible with Perl).
411 33. If --only-matching or --colour was set on a pcregrep call whose pattern
412 had alternative anchored branches, the search for a second match in a line
413 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
414 matched the line "0102" twice. The same bug affected patterns that started
415 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
416 twice.
418 34. Previously, PCRE did not allow quantification of assertions. However, Perl
419 does, and because of capturing effects, quantifying parenthesized
420 assertions may at times be useful. Quantifiers are now allowed for
421 parenthesized assertions.
423 35. A minor code tidy in pcre_compile() when checking options for \R usage.
425 36. \g was being checked for fancy things in a character class, when it should
426 just be a literal "g".
428 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
429 appearance of a nested POSIX class supersedes an apparent external class.
430 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
431 unescaped square brackets may also appear as part of class names. For
432 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
433 more like Perl. (But see 8.20/1 above.)
435 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
436 was because it thought it was \N{name}, which is not supported).
438 39. Add minix to OS list not supporting the -S option in pcretest.
440 40. PCRE tries to detect cases of infinite recursion at compile time, but it
441 cannot analyze patterns in sufficient detail to catch mutual recursions
442 such as ((?1))((?2)). There is now a runtime test that gives an error if a
443 subgroup is called recursively as a subpattern for a second time at the
444 same position in the subject string. In previous releases this might have
445 been caught by the recursion limit, or it might have run out of stack.
447 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
448 happen only once. PCRE was, however incorrectly giving a compile time error
449 "recursive call could loop indefinitely" because it cannot analyze the
450 pattern in sufficient detail. The compile time test no longer happens when
451 PCRE is compiling a conditional subpattern, but actual runaway loops are
452 now caught at runtime (see 40 above).
454 42. It seems that Perl allows any characters other than a closing parenthesis
455 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
456 has been changed to be the same.
458 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
459 as not to get warnings when autogen.sh is called. Also changed
460 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
462 44. To help people who use pcregrep to scan files containing exceedingly long
463 lines, the following changes have been made:
465 (a) The default value of the buffer size parameter has been increased from
466 8K to 20K. (The actual buffer used is three times this size.)
468 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
469 PCRE is built.
471 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
472 to be set at run time.
474 (d) Numerical values in pcregrep options can be followed by K or M, for
475 example --buffer-size=50K.
477 (e) If a line being scanned overflows pcregrep's buffer, an error is now
478 given and the return code is set to 2.
480 45. Add a pointer to the latest mark to the callout data block.
482 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
483 partial match of an empty string instead of no match. This was specific to
484 the use of ".".
486 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
487 complete match instead of a partial match. This bug was dependent on both
488 the PCRE_UTF8 and PCRE_DOTALL options being set.
490 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
491 starting byte set, because \b was not being ignored.
494 Version 8.12 15-Jan-2011
495 ------------------------
497 1. Fixed some typos in the markup of the man pages, and wrote a script that
498 checks for such things as part of the documentation building process.
500 2. On a big-endian 64-bit system, pcregrep did not correctly process the
501 --match-limit and --recursion-limit options (added for 8.11). In
502 particular, this made one of the standard tests fail. (The integer value
503 went into the wrong half of a long int.)
505 3. If the --colour option was given to pcregrep with -v (invert match), it
506 did strange things, either producing crazy output, or crashing. It should,
507 of course, ignore a request for colour when reporting lines that do not
508 match.
510 4. Another pcregrep bug caused similar problems if --colour was specified with
511 -M (multiline) and the pattern match finished with a line ending.
513 5. In pcregrep, when a pattern that ended with a literal newline sequence was
514 matched in multiline mode, the following line was shown as part of the
515 match. This seems wrong, so I have changed it.
517 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
518 the check for further matches in the same line (so they could be coloured)
519 to overrun the end of the current line. If another match was found, it was
520 incorrectly shown (and then shown again when found in the next line).
522 7. If pcregrep was compiled under Windows, there was a reference to the
523 function pcregrep_exit() before it was defined. I am assuming this was
524 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
525 reported by a user. I've moved the definition above the reference.
528 Version 8.11 10-Dec-2010
529 ------------------------
531 1. (*THEN) was not working properly if there were untried alternatives prior
532 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
533 backtracked to try for "b" instead of moving to the next alternative branch
534 at the same level (in this case, to look for "c"). The Perl documentation
535 is clear that when (*THEN) is backtracked onto, it goes to the "next
536 alternative in the innermost enclosing group".
538 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
539 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
540 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
541 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
542 (*THEN).
544 3. If \s appeared in a character class, it removed the VT character from
545 the class, even if it had been included by some previous item, for example
546 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
547 of \s, but is part of the POSIX "space" class.)
549 4. A partial match never returns an empty string (because you can always
550 match an empty string at the end of the subject); however the checking for
551 an empty string was starting at the "start of match" point. This has been
552 changed to the "earliest inspected character" point, because the returned
553 data for a partial match starts at this character. This means that, for
554 example, /(?<=abc)def/ gives a partial match for the subject "abc"
555 (previously it gave "no match").
557 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
558 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
559 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
560 has an implication that the given string is incomplete (because a partial
561 match is preferred over a full match). For this reason, these items now
562 give a partial match in this situation. [Aside: previously, the one case
563 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
564 match rather than a full match, which was wrong by the old rules, but is
565 now correct.]
567 6. There was a bug in the handling of #-introduced comments, recognized when
568 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
569 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
570 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
571 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
572 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
573 places in pcre_compile().
575 7. Related to (6) above, when pcre_compile() was skipping #-introduced
576 comments when looking ahead for named forward references to subpatterns,
577 the only newline sequence it recognized was NL. It now handles newlines
578 according to the set newline convention.
580 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
581 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
582 cater for a lack of strerror(). These oversights have been fixed.
584 9. Added --match-limit and --recursion-limit to pcregrep.
586 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
588 11. When the -o option was used, pcregrep was setting a return code of 1, even
589 when matches were found, and --line-buffered was not being honoured.
591 12. Added an optional parentheses number to the -o and --only-matching options
592 of pcregrep.
594 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
595 can match an empty string. The code to do it in pcretest and pcredemo
596 needed fixing:
598 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
599 only one byte after an empty string match just before CRLF (this case
600 just got forgotten; "any" and "anycrlf" were OK).
602 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
603 mode when an empty string match preceded an ASCII character followed by
604 a non-ASCII character. (The code for advancing by one character rather
605 than one byte was nonsense.)
607 (c) The pcredemo.c sample program did not have any code at all to handle
608 the cases when CRLF is a valid newline sequence.
610 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
611 as a starting offset was within the subject string. There is now a new
612 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
613 negative or greater than the length of the string. In order to test this,
614 pcretest is extended to allow the setting of negative starting offsets.
616 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
617 starting offset points to the beginning of a UTF-8 character was
618 unnecessarily clumsy. I tidied it up.
620 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
621 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
623 17. Nobody had reported that the --include_dir option, which was added in
624 release 7.7 should have been called --include-dir (hyphen, not underscore)
625 for compatibility with GNU grep. I have changed it to --include-dir, but
626 left --include_dir as an undocumented synonym, and the same for
627 --exclude-dir, though that is not available in GNU grep, at least as of
628 release 2.5.4.
630 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
631 characters from a string of bytes) have been redefined so as not to use
632 loops, in order to improve performance in some environments. At the same
633 time, I abstracted some of the common code into auxiliary macros to save
634 repetition (this should not affect the compiled code).
636 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
637 compile-time error is now given if \c is not followed by an ASCII
638 character, that is, a byte less than 128. (In EBCDIC mode, the code is
639 different, and any byte value is allowed.)
641 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
642 START_OPTIMIZE option, which is now allowed at compile time - but just
643 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
644 to pcregrep and other applications that have no direct access to PCRE
645 options. The new /Y option in pcretest sets this option when calling
646 pcre_compile().
648 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
649 back references. Groups containing recursive back references were forced to
650 be atomic by that change, but in the case of named groups, the amount of
651 memory required was incorrectly computed, leading to "Failed: internal
652 error: code overflow". This has been fixed.
654 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
655 pcretest.c, to avoid build problems in some Borland environments.
658 Version 8.10 25-Jun-2010
659 ------------------------
661 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
662 THEN.
664 2. (*ACCEPT) was not working when inside an atomic group.
666 3. Inside a character class, \B is treated as a literal by default, but
667 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
668 causes the error). The code is unchanged, but I tidied the documentation.
670 4. Inside a character class, PCRE always treated \R and \X as literals,
671 whereas Perl faults them if its -w option is set. I have changed PCRE so
672 that it faults them when PCRE_EXTRA is set.
674 5. Added support for \N, which always matches any character other than
675 newline. (It is the same as "." when PCRE_DOTALL is not set.)
677 6. When compiling pcregrep with newer versions of gcc which may have
678 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
679 declared with attribute warn_unused_result" were given. Just casting the
680 result to (void) does not stop the warnings; a more elaborate fudge is
681 needed. I've used a macro to implement this.
683 7. Minor change to pcretest.c to avoid a compiler warning.
685 8. Added four artifical Unicode properties to help with an option to make
686 \s etc use properties (see next item). The new properties are: Xan
687 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
689 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
690 use Unicode properties. (*UCP) at the start of a pattern can be used to set
691 this option. Modified pcretest to add /W to test this facility. Added
692 REG_UCP to make it available via the POSIX interface.
694 10. Added --line-buffered to pcregrep.
696 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
697 studied, and the match started with a letter with a code point greater than
698 127 whose first byte was different to the first byte of the other case of
699 the letter, the other case of this starting letter was not recognized
700 (#976).
702 12. If a pattern that was studied started with a repeated Unicode property
703 test, for example, \p{Nd}+, there was the theoretical possibility of
704 setting up an incorrect bitmap of starting bytes, but fortunately it could
705 not have actually happened in practice until change 8 above was made (it
706 added property types that matched character-matching opcodes).
708 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
709 possible starting bytes for non-anchored patterns.
711 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
712 \R, and also a number of cases that involve Unicode properties, both
713 explicit and implicit when PCRE_UCP is set.
715 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
716 input, it could crash or give wrong results if characters with values
717 greater than 0xc0 were present in the subject string. (Detail: it assumed
718 UTF-8 input when processing these items.)
720 16. Added a lot of (int) casts to avoid compiler warnings in systems where
721 size_t is 64-bit (#991).
723 17. Added a check for running out of memory when PCRE is compiled with
724 --disable-stack-for-recursion (#990).
726 18. If the last data line in a file for pcretest does not have a newline on
727 the end, a newline was missing in the output.
729 19. The default pcre_chartables.c file recognizes only ASCII characters (values
730 less than 128) in its various bitmaps. However, there is a facility for
731 generating tables according to the current locale when PCRE is compiled. It
732 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
733 characters, are recognized by isspace() and therefore were getting set in
734 these tables, and indeed these tables seem to approximate to ISO 8859. This
735 caused a problem in UTF-8 mode when pcre_study() was used to create a list
736 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
737 which of course cannot start UTF-8 characters. I have changed the code so
738 that only real ASCII characters (less than 128) and the correct starting
739 bytes for UTF-8 encodings are set for characters greater than 127 when in
740 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
741 altogether.)
743 20. Added the /T option to pcretest so as to be able to run tests with non-
744 standard character tables, thus making it possible to include the tests
745 used for 19 above in the standard set of tests.
747 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
748 reference to a subpattern the other side of a comment that contains an
749 opening parenthesis caused either an internal compiling error, or a
750 reference to the wrong subpattern.
753 Version 8.02 19-Mar-2010
754 ------------------------
756 1. The Unicode data tables have been updated to Unicode 5.2.0.
758 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
759 configured.
761 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
762 original author of that file, following a query about its status.
764 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
765 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
767 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
768 quantifier applied to a forward-referencing subroutine call, could compile
769 incorrect code or give the error "internal error: previously-checked
770 referenced subpattern not found".
772 6. Both MS Visual Studio and Symbian OS have problems with initializing
773 variables to point to external functions. For these systems, therefore,
774 pcre_malloc etc. are now initialized to local functions that call the
775 relevant global functions.
777 7. There were two entries missing in the vectors called coptable and poptable
778 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
779 I've fixed the data, and added a kludgy way of testing at compile time that
780 the lengths are correct (equal to the number of opcodes).
782 8. Following on from 7, I added a similar kludge to check the length of the
783 eint vector in pcreposix.c.
785 9. Error texts for pcre_compile() are held as one long string to avoid too
786 much relocation at load time. To find a text, the string is searched,
787 counting zeros. There was no check for running off the end of the string,
788 which could happen if a new error number was added without updating the
789 string.
791 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
793 11. \K was not working if it appeared in an atomic group or in a group that
794 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
795 \K is "not well defined" if used in an assertion. PCRE now accepts it if
796 the assertion is positive, but not if it is negative.
798 12. Change 11 fortuitously reduced the size of the stack frame used in the
799 "match()" function of pcre_exec.c by one pointer. Forthcoming
800 implementation of support for (*MARK) will need an extra pointer on the
801 stack; I have reserved it now, so that the stack frame size does not
802 decrease.
804 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
805 item in branch that calls a recursion is a subroutine call - as in the
806 second branch in the above example - was incorrectly given the compile-
807 time error "recursive call could loop indefinitely" because pcre_compile()
808 was not correctly checking the subroutine for matching a non-empty string.
810 14. The checks for overrunning compiling workspace could trigger after an
811 overrun had occurred. This is a "should never occur" error, but it can be
812 triggered by pathological patterns such as hundreds of nested parentheses.
813 The checks now trigger 100 bytes before the end of the workspace.
815 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
818 Version 8.01 19-Jan-2010
819 ------------------------
821 1. If a pattern contained a conditional subpattern with only one branch (in
822 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
823 computed the wrong minimum data length (which is of course zero for such
824 subpatterns). This could cause incorrect "no match" results.
826 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
827 the pattern is reset in the first branch, pcre_compile() failed with
828 "internal error: code overflow at offset...". This happened only when
829 the reset was to the original external option setting. (An optimization
830 abstracts leading options settings into an external setting, which was the
831 cause of this.)
833 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
834 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
835 assertion pattern did not match (meaning that the assertion was true), it
836 was incorrectly treated as false if the SKIP had been reached during the
837 matching. This also applied to assertions used as conditions.
839 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
840 assertion subpattern, including such a pattern used as a condition,
841 unpredictable results occurred, instead of the error return
844 5. The C++ GlobalReplace function was not working like Perl for the special
845 situation when an empty string is matched. It now does the fancy magic
846 stuff that is necessary.
848 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
849 removed. (These were left over from very, very early versions of PCRE.)
851 7. Some cosmetic changes to the code to make life easier when compiling it
852 as part of something else:
854 (a) Change DEBUG to PCRE_DEBUG.
856 (b) In pcre_compile(), rename the member of the "branch_chain" structure
857 called "current" as "current_branch", to prevent a collision with the
858 Linux macro when compiled as a kernel module.
860 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
861 prevent a collision with the Linux macro when compiled as a kernel
862 module.
864 8. In pcre_compile() there are some checks for integer overflows that used to
865 cast potentially large values to (double). This has been changed to that
866 when building, a check for int64_t is made, and if it is found, it is used
867 instead, thus avoiding the use of floating point arithmetic. (There is no
868 other use of FP in PCRE.) If int64_t is not found, the fallback is to
869 double.
871 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
872 2005 (difference between two addresses compared to an unsigned value).
874 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
875 custom one, because of the following reported problem in Windows:
877 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
878 under Win32.
879 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
880 therefore missing the function definition.
881 - The compiler thus generates a "C" signature for the test function.
882 - The linker fails to find the "C" function.
883 - PCRE fails to configure if asked to do so against libbz2.
885 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
886 messages were output:
888 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
889 rerunning libtoolize, to keep the correct libtool macros in-tree.
890 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
892 I have done both of these things.
894 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
895 most of the time, it *can* run out if it is given a pattern that contains a
896 runaway infinite recursion. I updated the discussion in the pcrestack man
897 page.
899 13. Now that we have gone to the x.xx style of version numbers, the minor
900 version may start with zero. Using 08 or 09 is a bad idea because users
901 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
902 interpreted as invalid octal numbers. I've updated the previous comment in
903 configure.ac, and also added a check that gives an error if 08 or 09 are
904 used.
906 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
907 causing partial matching to fail when the end of the subject matched \W
908 in a UTF-8 pattern where \W was quantified with a minimum of 3.
910 15. There were some discrepancies between the declarations in pcre_internal.h
911 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
912 their definitions. The declarations used "const uschar *" and the
913 definitions used USPTR. Even though USPTR is normally defined as "const
914 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
915 reported that: "This difference in casting confuses some C++ compilers, for
916 example, SunCC recognizes above declarations as different functions and
917 generates broken code for hbpcre." I have changed the declarations to use
918 USPTR.
920 16. GNU libtool is named differently on some systems. The autogen.sh script now
921 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
922 (FreeBSD).
924 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
925 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
926 comment: "Figure out how to create a longlong from a string: strtoll and
927 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
928 instance, but it only takes 2 args instead of 3!"
930 18. A subtle bug concerned with back references has been fixed by a change of
931 specification, with a corresponding code fix. A pattern such as
932 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
933 refers, was giving matches when it shouldn't. For example, xa=xaaa would
934 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
935 same bug. Such groups have to be quantified to be useful, or contained
936 inside another quantified group. (If there's no repetition, the reference
937 can never match.) The problem arises because, having left the group and
938 moved on to the rest of the pattern, a later failure that backtracks into
939 the group uses the captured value from the final iteration of the group
940 rather than the correct earlier one. I have fixed this in PCRE by forcing
941 any group that contains a reference to itself to be an atomic group; that
942 is, there cannot be any backtracking into it once it has completed. This is
943 similar to recursive and subroutine calls.
946 Version 8.00 19-Oct-09
947 ----------------------
949 1. The table for translating pcre_compile() error codes into POSIX error codes
950 was out-of-date, and there was no check on the pcre_compile() error code
951 being within the table. This could lead to an OK return being given in
952 error.
954 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
955 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
956 in a Windows environment.
958 3. The pcregrep --count option prints the count for each file even when it is
959 zero, as does GNU grep. However, pcregrep was also printing all files when
960 --files-with-matches was added. Now, when both options are given, it prints
961 counts only for those files that have at least one match. (GNU grep just
962 prints the file name in this circumstance, but including the count seems
963 more useful - otherwise, why use --count?) Also ensured that the
964 combination -clh just lists non-zero counts, with no names.
966 4. The long form of the pcregrep -F option was incorrectly implemented as
967 --fixed_strings instead of --fixed-strings. This is an incompatible change,
968 but it seems right to fix it, and I didn't think it was worth preserving
969 the old behaviour.
971 5. The command line items --regex=pattern and --regexp=pattern were not
972 recognized by pcregrep, which required --regex pattern or --regexp pattern
973 (with a space rather than an '='). The man page documented the '=' forms,
974 which are compatible with GNU grep; these now work.
976 6. No libpcreposix.pc file was created for pkg-config; there was just
977 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
979 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
980 when UCP support is not needed, by modifying the Python script that
981 generates it from Unicode data files. This should not matter if the module
982 is correctly used as a library, but I received one complaint about 50K of
983 unwanted data. My guess is that the person linked everything into his
984 program rather than using a library. Anyway, it does no harm.
986 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
987 was a minimum greater than 1 for a wide character in a possessive
988 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
989 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
990 character. Chaos in the form of incorrect output or a compiling loop could
991 result.
993 9. The restrictions on what a pattern can contain when partial matching is
994 requested for pcre_exec() have been removed. All patterns can now be
995 partially matched by this function. In addition, if there are at least two
996 slots in the offset vector, the offset of the earliest inspected character
997 for the match and the offset of the end of the subject are set in them when
998 PCRE_ERROR_PARTIAL is returned.
1000 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1001 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1002 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1003 and may be more useful for multi-segment matching.
1005 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1006 used to be given if ever the end of the subject was reached; now it is
1007 given only if matching could not proceed because another character was
1008 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1009 string "Z", which now yields "no match" instead of "partial match". In the
1010 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1011 final character ended with (*FAIL).
1013 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1014 if the pattern had a "must contain" character that was already found in the
1015 earlier partial match, unless partial matching was again requested. For
1016 example, with the pattern /dog.(body)?/, the "must contain" character is
1017 "g". If the first part-match was for the string "dog", restarting with
1018 "sbody" failed. This bug has been fixed.
1020 13. The string returned by pcre_dfa_exec() after a partial match has been
1021 changed so that it starts at the first inspected character rather than the
1022 first character of the match. This makes a difference only if the pattern
1023 starts with a lookbehind assertion or \b or \B (\K is not supported by
1024 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1025 matching functions compatible, and I think it's the right thing to do.
1027 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1028 so that the demonstration program is easily available in environments where
1029 PCRE has not been installed from source.
1031 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1032 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1033 library.
1035 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1036 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1037 is not the first non-POSIX option to be added. Clearly some people find
1038 these options useful.
1040 17. If a caller to the POSIX matching function regexec() passes a non-zero
1041 value for nmatch with a NULL value for pmatch, the value of
1042 nmatch is forced to zero.
1044 18. RunGrepTest did not have a test for the availability of the -u option of
1045 the diff command, as RunTest does. It now checks in the same way as
1046 RunTest, and also checks for the -b option.
1048 19. If an odd number of negated classes containing just a single character
1049 interposed, within parentheses, between a forward reference to a named
1050 subpattern and the definition of the subpattern, compilation crashed with
1051 an internal error, complaining that it could not find the referenced
1052 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1053 [The bug was that it was starting one character too far in when skipping
1054 over the character class, thus treating the ] as data rather than
1055 terminating the class. This meant it could skip too much.]
1057 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1058 /g option in pcretest when the pattern contains \K, which makes it possible
1059 to have an empty string match not at the start, even when the pattern is
1060 anchored. Updated pcretest and pcredemo to use this option.
1062 21. If the maximum number of capturing subpatterns in a recursion was greater
1063 than the maximum at the outer level, the higher number was returned, but
1064 with unset values at the outer level. The correct (outer level) value is
1065 now given.
1067 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1068 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1069 make it do so. The string so far is captured, making this feature
1070 compatible with Perl.
1072 23. The tests have been re-organized, adding tests 11 and 12, to make it
1073 possible to check the Perl 5.10 features against Perl 5.10.
1075 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1076 pattern matches a fixed length string. PCRE did not allow this; now it
1077 does. Neither allows recursion.
1079 25. I finally figured out how to implement a request to provide the minimum
1080 length of subject string that was needed in order to match a given pattern.
1081 (It was back references and recursion that I had previously got hung up
1082 on.) This code has now been added to pcre_study(); it finds a lower bound
1083 to the length of subject needed. It is not necessarily the greatest lower
1084 bound, but using it to avoid searching strings that are too short does give
1085 some useful speed-ups. The value is available to calling programs via
1086 pcre_fullinfo().
1088 26. While implementing 25, I discovered to my embarrassment that pcretest had
1089 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1090 study optimizations had never been tested with that matching function.
1091 Oops. What is worse, even when it was passed study data, there was a bug in
1092 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1093 were also very few tests of studied patterns with pcre_dfa_exec().
1095 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1096 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1097 on the other side of the coin, they are no longer allowed to have different
1098 names, because these cannot be distinguished in PCRE, and this has caused
1099 confusion. (This is a difference from Perl.)
1101 28. When duplicate subpattern names are present (necessarily with different
1102 numbers, as required by 27 above), and a test is made by name in a
1103 conditional pattern, either for a subpattern having been matched, or for
1104 recursion in such a pattern, all the associated numbered subpatterns are
1105 tested, and the overall condition is true if the condition is true for any
1106 one of them. This is the way Perl works, and is also more like the way
1107 testing by number works.
1110 Version 7.9 11-Apr-09
1111 ---------------------
1113 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1114 (pcretest), all targets were linked against these libraries. This included
1115 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1116 libraries. This caused unwanted dependencies to be created. This problem
1117 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1118 pcretest is linked with readline.
1120 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1121 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1122 moved outside it again, because FALSE and TRUE are already defined in AIX,
1123 but BOOL is not.
1125 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1126 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1128 4. The pcregrep documentation said spaces were inserted as well as colons (or
1129 hyphens) following file names and line numbers when outputting matching
1130 lines. This is not true; no spaces are inserted. I have also clarified the
1131 wording for the --colour (or --color) option.
1133 5. In pcregrep, when --colour was used with -o, the list of matching strings
1134 was not coloured; this is different to GNU grep, so I have changed it to be
1135 the same.
1137 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1138 each matching line was coloured. Now it goes on to look for further matches
1139 of any of the test patterns, which is the same behaviour as GNU grep.
1141 7. A pattern that could match an empty string could cause pcregrep to loop; it
1142 doesn't make sense to accept an empty string match in pcregrep, so I have
1143 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1144 seems to be how GNU grep behaves.
1146 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1147 start or after a newline", because the conditional assertion was not being
1148 correctly handled. The rule now is that both the assertion and what follows
1149 in the first alternative must satisfy the test.
1151 9. If auto-callout was enabled in a pattern with a conditional group whose
1152 condition was an assertion, PCRE could crash during matching, both with
1153 pcre_exec() and pcre_dfa_exec().
1155 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1156 used for matching.
1158 11. Unicode property support in character classes was not working for
1159 characters (bytes) greater than 127 when not in UTF-8 mode.
1161 12. Added the -M command line option to pcretest.
1163 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1165 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1167 16. Added comments and documentation about mis-use of no_arg in the C++
1168 wrapper.
1170 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1171 from Martin Jerabek that uses macro names for all relevant character and
1172 string constants.
1174 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1175 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1176 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1177 these, but not everybody uses configure.
1179 19. A conditional group that had only one branch was not being correctly
1180 recognized as an item that could match an empty string. This meant that an
1181 enclosing group might also not be so recognized, causing infinite looping
1182 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1183 with the subject "ab", where knowledge that the repeated group can match
1184 nothing is needed in order to break the loop.
1186 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1187 exec(), but without supplying a callout function, matching went wrong.
1189 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1190 leak if the size of the offset vector was greater than 30. When the vector
1191 is smaller, the saved offsets during recursion go onto a local stack
1192 vector, but for larger vectors malloc() is used. It was failing to free
1193 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1194 error, in fact).
1196 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1197 heapframe that is used only when UTF-8 support is enabled. This caused no
1198 problem, but was untidy.
1200 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1201 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1202 included within another project.
1204 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1205 slightly modified by me:
1207 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1208 not building pcregrep.
1210 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1211 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1213 25. Forward references, both numeric and by name, in patterns that made use of
1214 duplicate group numbers, could behave incorrectly or give incorrect errors,
1215 because when scanning forward to find the reference group, PCRE was not
1216 taking into account the duplicate group numbers. A pattern such as
1217 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1219 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1220 the feature of a custom pointer more persuasive (as requested by a user).
1222 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1223 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1224 pre-defined.
1226 28. Added support for (*UTF8) at the start of a pattern.
1228 29. Arrange for flags added by the "release type" setting in CMake to be shown
1229 in the configuration summary.
1232 Version 7.8 05-Sep-08
1233 ---------------------
1235 1. Replaced UCP searching code with optimized version as implemented for Ad
1236 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1237 stage table and inline lookup instead of a function, giving speed ups of 2
1238 to 5 times on some simple patterns that I tested. Permission was given to
1239 distribute the MultiStage2.py script that generates the tables (it's not in
1240 the tarball, but is in the Subversion repository).
1242 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1243 scripts.
1245 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1246 a group with a zero qualifier. The result of the study could be incorrect,
1247 or the function might crash, depending on the pattern.
1249 4. Caseless matching was not working for non-ASCII characters in back
1250 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1251 It now works when Unicode Property Support is available.
1253 5. In pcretest, an escape such as \x{de} in the data was always generating
1254 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1255 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1256 truncation.
1258 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1260 7. Added two (int) casts to pcregrep when printing the difference of two
1261 pointers, in case they are 64-bit values.
1263 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1264 test 2 if it fails.
1266 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1267 and a #define of that name to empty if it is not externally set. This is to
1268 allow users of MSVC to set it if necessary.
1270 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1271 the convenience functions in the pcre_get.c source file.
1273 11. An option change at the start of a pattern that had top-level alternatives
1274 could cause overwriting and/or a crash. This command provoked a crash in
1275 some environments:
1277 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1279 This potential security problem was recorded as CVE-2008-2371.
1281 12. For a pattern where the match had to start at the beginning or immediately
1282 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1283 pcre_dfa_exec() could read past the end of the passed subject if there was
1284 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1285 pcretest so that it places the subject at the end of its malloc-ed buffer.
1287 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1288 exec() might read past the end of the data buffer in UTF-8 mode.
1290 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1291 the data contained the byte 0x85 as part of a UTF-8 character within its
1292 first line. This applied both to normal and DFA matching.
1294 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1295 /^[^d]*?$/8 failed to match "abc".
1297 16. Added a missing copyright notice to pcrecpp_internal.h.
1299 17. Make it more clear in the documentation that values returned from
1300 pcre_exec() in ovector are byte offsets, not character counts.
1302 18. Tidied a few places to stop certain compilers from issuing warnings.
1304 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1305 supplied by Stefan Weber. I made a further small update for 7.8 because
1306 there is a change of source arrangements: the pcre_searchfuncs.c module is
1307 replaced by pcre_ucd.c.
1310 Version 7.7 07-May-08
1311 ---------------------
1313 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1314 a string to a long long, pretend we don't even have a long long." This is
1315 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1317 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1318 pre-7.6 versions, which defined a global no_arg variable instead of putting
1319 it in the RE class. (See also #8 below.)
1321 3. Remove a line of dead code, identified by coverity and reported by Nuno
1322 Lopes.
1324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1326 (1) The include/exclude patterns were being applied to the whole pathnames
1327 of files, instead of just to the final components.
1329 (2) If there was more than one level of directory, the subdirectories were
1330 skipped unless they satisfied the include/exclude conditions. This is
1331 inconsistent with GNU grep (and could even be seen as contrary to the
1332 pcregrep specification - which I improved to make it absolutely clear).
1333 The action now is always to scan all levels of directory, and just
1334 apply the include/exclude patterns to regular files.
1336 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1337 --exclude_dir in the tests to avoid scanning .svn directories.
1339 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1340 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1341 doesn't support NULs in patterns.
1343 7. Added some missing "const"s to declarations of static tables in
1344 pcre_compile.c and pcre_dfa_exec.c.
1346 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1347 caused by fix #2 above. (Subsequently also a second patch to fix the
1348 first patch. And a third patch - this was a messy problem.)
1350 9. Applied Craig's patch to remove the use of push_back().
1352 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1353 matching function regexec().
1355 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1356 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1357 references. PCRE supports relative numbers with this syntax (I don't think
1358 Oniguruma does).
1360 12. Previously, a group with a zero repeat such as (...){0} was completely
1361 omitted from the compiled regex. However, this means that if the group
1362 was called as a subroutine from elsewhere in the pattern, things went wrong
1363 (an internal error was given). Such groups are now left in the compiled
1364 pattern, with a new opcode that causes them to be skipped at execution
1365 time.
1367 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1368 to the way PCRE behaves:
1370 (a) A lone ] character is dis-allowed (Perl treats it as data).
1372 (b) A back reference to an unmatched subpattern matches an empty string
1373 (Perl fails the current match path).
1375 (c) A data ] in a character class must be notated as \] because if the
1376 first data character in a class is ], it defines an empty class. (In
1377 Perl it is not possible to have an empty class.) The empty class []
1378 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1379 The negative empty class [^] matches any one character, independently
1380 of the DOTALL setting.
1382 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1383 non-existent subpattern following a character class starting with ']' and
1384 containing () gave an internal compiling error instead of "reference to
1385 non-existent subpattern". Fortunately, when the pattern did exist, the
1386 compiled code was correct. (When scanning forwards to check for the
1387 existencd of the subpattern, it was treating the data ']' as terminating
1388 the class, so got the count wrong. When actually compiling, the reference
1389 was subsequently set up correctly.)
1391 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1392 it was being rejected as not supported by pcre_dfa_exec(), even though
1393 other assertions are supported. I have made pcre_dfa_exec() support
1394 (*FAIL).
1396 16. The implementation of 13c above involved the invention of a new opcode,
1397 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1398 cannot be changed at match time, I realized I could make a small
1399 improvement to matching performance by compiling OP_ALLANY instead of
1400 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1401 on the OP_ANY path.
1403 17. Compiling pcretest on Windows with readline support failed without the
1404 following two fixes: (1) Make the unistd.h include conditional on
1405 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1407 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1408 ncurses library to be included for pcretest when ReadLine support is
1409 requested, but also to allow for it to be overridden. This patch came from
1410 Daniel Bergström.
1412 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1413 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1414 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1415 spotting this.
1418 Version 7.6 28-Jan-08
1419 ---------------------
1421 1. A character class containing a very large number of characters with
1422 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1423 overflow.
1425 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1426 HAVE_LONG_LONG is not defined.
1428 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1429 bring them up to date and include new features. This patch includes:
1431 - Fixed PH's badly added libz and libbz2 support.
1432 - Fixed a problem with static linking.
1433 - Added pcredemo. [But later removed - see 7 below.]
1434 - Fixed dftables problem and added an option.
1435 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1437 - Added readline support for pcretest.
1438 - Added an listing of the option settings after cmake has run.
1440 4. A user submitted a patch to Makefile that makes it easy to create
1441 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1442 Makefile.am that cause it to include this special target, without
1443 affecting anything else. Note that the same mingw target plus all
1444 the other distribution libraries and programs are now supported
1445 when configuring with CMake (see 6 below) instead of with
1446 Configure/Make.
1448 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1449 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1450 exported in the Windows port". It has not yet been confirmed that the patch
1451 solves the problem, but it does no harm.
1453 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1454 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1455 with CMake, and also correct the comment about stack recursion.
1457 7. Remove the automatic building of pcredemo from the ./configure system and
1458 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1459 of a program that users should build themselves after PCRE is installed, so
1460 building it automatically is not really right. What is more, it gave
1461 trouble in some build environments.
1463 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1466 Version 7.5 10-Jan-08
1467 ---------------------
1469 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1470 values in parens when parsing an RE using the C++ wrapper."
1472 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1473 Characters greater than 255 were excluded from the class instead of being
1474 included.
1476 3. The same bug as (2) above applied to negated POSIX classes such as
1477 [:^space:].
1479 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1480 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1481 I have changed it.
1483 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1484 first named subpattern) and a construct such as (?&a) would reference the
1485 first named subpattern whose name started with "a" (in other words, the
1486 length check was missing). Both these problems are fixed. "Subpattern name
1487 expected" is now given for (?&) (a zero-length name), and this patch also
1488 makes it give the same error for \k'' (previously it complained that that
1489 was a reference to a non-existent subpattern).
1491 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1492 this is right because (?- can be followed by option settings as well as by
1493 digits. I have, however, made the messages clearer.
1495 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1496 than the number used in the conditional) now cause a compile-time error.
1497 This is actually not compatible with Perl, which accepts such patterns, but
1498 treats the conditional as always being FALSE (as PCRE used to), but it
1499 seems to me that giving a diagnostic is better.
1501 8. Change "alphameric" to the more common word "alphanumeric" in comments
1502 and messages.
1504 9. Fix two occurrences of "backslash" in comments that should have been
1505 "backspace".
1507 10. Remove two redundant lines of code that can never be obeyed (their function
1508 was moved elsewhere).
1510 11. The program that makes PCRE's Unicode character property table had a bug
1511 which caused it to generate incorrect table entries for sequences of
1512 characters that have the same character type, but are in different scripts.
1513 It amalgamated them into a single range, with the script of the first of
1514 them. In other words, some characters were in the wrong script. There were
1515 thirteen such cases, affecting characters in the following ranges:
1517 U+002b0 - U+002c1
1518 U+0060c - U+0060d
1519 U+0061e - U+00612
1520 U+0064b - U+0065e
1521 U+0074d - U+0076d
1522 U+01800 - U+01805
1523 U+01d00 - U+01d77
1524 U+01d9b - U+01dbf
1525 U+0200b - U+0200f
1526 U+030fc - U+030fe
1527 U+03260 - U+0327f
1528 U+0fb46 - U+0fbb1
1529 U+10450 - U+1049d
1531 12. The -o option (show only the matching part of a line) for pcregrep was not
1532 compatible with GNU grep in that, if there was more than one match in a
1533 line, it showed only the first of them. It now behaves in the same way as
1534 GNU grep.
1536 13. If the -o and -v options were combined for pcregrep, it printed a blank
1537 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1538 does the same. The return code can be used to tell if there were any
1539 non-matching lines.
1541 14. Added --file-offsets and --line-offsets to pcregrep.
1543 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1544 infinitely looping recursion. The bug was that positive lookaheads were not
1545 being skipped when checking for a possible empty match (negative lookaheads
1546 and both kinds of lookbehind were skipped).
1548 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1549 inclusion of <windows.h> to before rather than after the definition of
1550 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1552 17. Specifying a possessive quantifier with a specific limit for a Unicode
1553 character property caused pcre_compile() to compile bad code, which led at
1554 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1555 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1556 caused the error; without that there was no problem.
1558 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1560 19. Added --enable-pcretest-libreadline.
1562 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1563 RE::GlobalReplace(). As a result, the number of replacements returned was
1564 double what it should be. I removed one of the increments, but Craig sent a
1565 later patch that removed the other one (the right fix) and added unit tests
1566 that check the return values (which was not done before).
1568 21. Several CMake things:
1570 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1571 the names libpcre and libpcreposix, not just pcre and pcreposix.
1573 (2) The above change means that pcretest and pcregrep are now correctly
1574 linked with the newly-built libraries, not previously installed ones.
1578 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1579 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1580 UTF-8 newline character). The key issue is that the pattern starts .*;
1581 this means that the match must be either at the beginning, or after a
1582 newline. The bug was in the code for advancing after a failed match and
1583 checking that the new position followed a newline. It was not taking
1584 account of UTF-8 characters correctly.
1586 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1587 character classes. PCRE was not treating the sequence [:...:] as a
1588 character class unless the ... were all letters. Perl, however, seems to
1589 allow any characters between [: and :], though of course it rejects as
1590 unknown any "names" that contain non-letters, because all the known class
1591 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1592 for example, whereas PCRE did not - it did not recognize a POSIX character
1593 class. This seemed a bit dangerous, so the code has been changed to be
1594 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1595 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1596 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1597 Perl does, and where it didn't before.
1599 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1600 Windows environments %n is disabled by default.
1603 Version 7.4 21-Sep-07
1604 ---------------------
1606 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1607 means that a class such as [\s] counted as "explicit reference to CR or
1608 LF". That isn't really right - the whole point of the change was to try to
1609 help when there was an actual mention of one of the two characters. So now
1610 the change happens only if \r or \n (or a literal CR or LF) character is
1611 encountered.
1613 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1614 of both had grown to the point where there were only 3 bits left.
1615 Fortunately, there was spare space in the data structure, and so I have
1616 moved the internal flags into a new 16-bit field to free up more option
1617 bits.
1619 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1620 but did not set the internal JCHANGED flag - either of these is enough to
1621 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1622 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1623 start sets both bits.
1625 4. Added options (at build time, compile time, exec time) to change \R from
1626 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1628 5. doc/pcresyntax.html was missing from the distribution.
1630 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1631 compatibility, even though it is no longer used.
1633 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1634 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1635 windows.h file is present (where different names are used). [This was
1636 reversed later after testing - see 16 below.]
1638 8. Changed all #include <config.h> to #include "config.h". There were also
1639 some further <pcre.h> cases that I changed to "pcre.h".
1641 9. When pcregrep was used with the --colour option, it missed the line ending
1642 sequence off the lines that it output.
1644 10. It was pointed out to me that arrays of string pointers cause lots of
1645 relocations when a shared library is dynamically loaded. A technique of
1646 using a single long string with a table of offsets can drastically reduce
1647 these. I have refactored PCRE in four places to do this. The result is
1648 dramatic:
1650 Originally: 290
1651 After changing UCP table: 187
1652 After changing error message table: 43
1653 After changing table of "verbs" 36
1654 After changing table of Posix names 22
1656 Thanks to the folks working on Gregex for glib for this insight.
1658 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1659 unicode-properties was also set.
1661 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1663 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1664 checked only for CRLF.
1666 14. Added casts to pcretest.c to avoid compiler warnings.
1668 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1670 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1671 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1672 entirely. This removes changes made in 7 above.
1674 17. The CMake files have been updated, and there is now more information about
1675 building with CMake in the NON-UNIX-USE document.
1678 Version 7.3 28-Aug-07
1679 ---------------------
1681 1. In the rejigging of the build system that eventually resulted in 7.1, the
1682 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1683 brackets there is not right, since it causes compilers to look for an
1684 installed pcre.h, not the version that is in the source that is being
1685 compiled (which of course may be different). I have changed it back to:
1687 #include "pcre.h"
1689 I have a vague recollection that the change was concerned with compiling in
1690 different directories, but in the new build system, that is taken care of
1691 by the VPATH setting the Makefile.
1693 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1694 when the subject happened to end in the byte 0x85 (e.g. if the last
1695 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1696 characters but of course it shouldn't be taken as a newline when it is part
1697 of another character. The bug was that, for an unlimited repeat of . in
1698 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1699 characters when looking for a newline.
1701 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1703 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1704 in debug output.
1706 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1707 long printing in the pcrecpp unittest when running under MinGW.
1709 6. ESC_K was left out of the EBCDIC table.
1711 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1712 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1713 limit also applies to "virtual nesting" when a pattern is recursive, and in
1714 this case 1000 isn't so big. I have been able to remove this limit at the
1715 expense of backing off one optimization in certain circumstances. Normally,
1716 when pcre_exec() would call its internal match() function recursively and
1717 immediately return the result unconditionally, it uses a "tail recursion"
1718 feature to save stack. However, when a subpattern that can match an empty
1719 string has an unlimited repetition quantifier, it no longer makes this
1720 optimization. That gives it a stack frame in which to save the data for
1721 checking that an empty string has been matched. Previously this was taken
1722 from the 1000-entry workspace that had been reserved. So now there is no
1723 explicit limit, but more stack is used.
1725 8. Applied Daniel's patches to solve problems with the import/export magic
1726 syntax that is required for Windows, and which was going wrong for the
1727 pcreposix and pcrecpp parts of the library. These were overlooked when this
1728 problem was solved for the main library.
1730 9. There were some crude static tests to avoid integer overflow when computing
1731 the size of patterns that contain repeated groups with explicit upper
1732 limits. As the maximum quantifier is 65535, the maximum group length was
1733 set at 30,000 so that the product of these two numbers did not overflow a
1734 32-bit integer. However, it turns out that people want to use groups that
1735 are longer than 30,000 bytes (though not repeat them that many times).
1736 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1737 made it possible to implement the integer overflow checks in a much more
1738 dynamic way, which I have now done. The artificial limitation on group
1739 length has been removed - we now have only the limit on the total length of
1740 the compiled pattern, which depends on the LINK_SIZE setting.
1742 10. Fixed a bug in the documentation for get/copy named substring when
1743 duplicate names are permitted. If none of the named substrings are set, the
1744 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1745 empty string.
1747 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1748 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1749 because the ] is interpreted as the first data character and the
1750 terminating ] is not found. PCRE has been made compatible with Perl in this
1751 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1752 cause memory overwriting.
1754 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1755 string has been matched (to stop an infinite loop). It was not recognizing
1756 a conditional subpattern that could match an empty string if that
1757 subpattern was within another subpattern. For example, it looped when
1758 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1759 condition was not nested. This bug has been fixed.
1761 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1762 past the start of the subject in the presence of bytes with the top bit
1763 set, for example "\x8aBCD".
1765 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1766 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1768 14. Optimized (?!) to (*FAIL).
1770 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1771 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1772 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1773 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1774 does: it's just the validity check that is more restrictive.
1776 16. Inserted checks for integer overflows during escape sequence (backslash)
1777 processing, and also fixed erroneous offset values for syntax errors during
1778 backslash processing.
1780 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1781 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1783 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1784 caused an overrun.
1786 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1787 something other than just ASCII characters) inside a group that had an
1788 unlimited repeat caused a loop at compile time (while checking to see
1789 whether the group could match an empty string).
1791 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1792 [\P{Any}] did so. (Error in the code for printing property names.)
1794 21. An orphan \E inside a character class could cause a crash.
1796 22. A repeated capturing bracket such as (A)? could cause a wild memory
1797 reference during compilation.
1799 23. There are several functions in pcre_compile() that scan along a compiled
1800 expression for various reasons (e.g. to see if it's fixed length for look
1801 behind). There were bugs in these functions when a repeated \p or \P was
1802 present in the pattern. These operators have additional parameters compared
1803 with \d, etc, and these were not being taken into account when moving along
1804 the compiled data. Specifically:
1806 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1807 length.
1809 (b) An item such as \pL+ within a repeated group could cause crashes or
1810 loops.
1812 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1813 "reference to non-existent subpattern" error.
1815 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1817 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1818 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1820 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1822 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1823 character were causing crashes (broken optimization).
1825 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1826 \p or \P) caused a compile-time loop.
1828 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1829 break. For example, the unstudied pattern [\r\n]A does not match the string
1830 "\r\nA" because change 7.0/46 below moves the current point on by two
1831 characters after failing to match at the start. However, the pattern \nA
1832 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1833 the same is true. There doesn't seem any very clean way out of this, but
1834 what I have chosen to do makes the common cases work: PCRE now takes note
1835 of whether there can be an explicit match for \r or \n anywhere in the
1836 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1837 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1838 pattern has explicit CR or LF references.
1840 29. Added (*CR) etc for changing newline setting at start of pattern.
1843 Version 7.2 19-Jun-07
1844 ---------------------
1846 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1847 which is apparently normally available under Windows.
1849 2. Re-jig the pcregrep tests with different newline settings in an attempt
1850 to make them independent of the local environment's newline setting.
1852 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1854 4. Some of the "internals" tests were previously cut out when the link size
1855 was not 2, because the output contained actual offsets. The recent new
1856 "Z" feature of pcretest means that these can be cut out, making the tests
1857 usable with all link sizes.
1859 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1860 stack recursion. This gives a massive performance boost under BSD, but just
1861 a small improvement under Linux. However, it saves one field in the frame
1862 in all cases.
1864 6. Added more features from the forthcoming Perl 5.10:
1866 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1867 recursion call. It refers to the nth most recently opened parentheses.
1869 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1870 to be opened parentheses.
1872 (c) Conditions that refer to capturing parentheses can be specified
1873 relatively, for example, (?(-2)... or (?(+3)...
1875 (d) \K resets the start of the current match so that everything before
1876 is not part of it.
1878 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1880 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1881 reference syntax.
1883 (g) (?| introduces a group in which the numbering of parentheses in each
1884 alternative starts with the same number.
1886 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1888 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1891 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1892 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1893 for detecting groups that can match an empty string.
1895 9. A pattern with a very large number of alternatives (more than several
1896 hundred) was running out of internal workspace during the pre-compile
1897 phase, where pcre_compile() figures out how much memory will be needed. A
1898 bit of new cunning has reduced the workspace needed for groups with
1899 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1900 workspace instead of running out of the 4096 that are available.
1902 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1904 11. Applied patch from Google to remove an optimization that didn't quite work.
1905 The report of the bug said:
1907 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1908 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1909 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1911 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1912 it matched the wrong number of bytes.
1915 Version 7.1 24-Apr-07
1916 ---------------------
1918 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1919 that is more "standard", making use of automake and other Autotools. There
1920 is some re-arrangement of the files and adjustment of comments consequent
1921 on this.
1923 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1924 for recursive directory scanning broke on some systems because the files
1925 are not scanned in any specific order and on different systems the order
1926 was different. A call to "sort" has been inserted into RunGrepTest for the
1927 approprate test as a short-term fix. In the longer term there may be an
1928 alternative.
1930 3. I had an email from Eric Raymond about problems translating some of PCRE's
1931 man pages to HTML (despite the fact that I distribute HTML pages, some
1932 people do their own conversions for various reasons). The problems
1933 concerned the use of low-level troff macros .br and .in. I have therefore
1934 removed all such uses from the man pages (some were redundant, some could
1935 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1936 HTML has been updated to handle .nf/.fi and to complain if it encounters
1937 .br or .in.
1939 4. Updated comments in configure.ac that get placed in config.h.in and also
1940 arranged for config.h to be included in the distribution, with the name
1941 config.h.generic, for the benefit of those who have to compile without
1942 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1944 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1945 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1946 makevp.bat for the current PCRE, using the additional files
1947 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1949 6. A Windows user reported a minor discrepancy with test 2, which turned out
1950 to be caused by a trailing space on an input line that had got lost in his
1951 copy. The trailing space was an accident, so I've just removed it.
1953 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1954 that is needed.
1956 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1957 as "const" (a) because they are and (b) because it helps the PHP
1958 maintainers who have recently made a script to detect big data structures
1959 in the php code that should be moved to the .rodata section. I remembered
1960 to update Builducptable as well, so it won't revert if ucptable.h is ever
1961 re-created.
1963 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1964 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1965 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1966 support is not required. This saves 1.5-2K of code, which is important in
1967 some applications.
1969 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1970 so as not to refer to the tables, even though these functions will never be
1971 called when UTF-8 support is disabled. Otherwise there are problems with a
1972 shared library.
1974 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1976 (a) It was defining its arguments as char * instead of void *.
1978 (b) It was assuming that all moves were upwards in memory; this was true
1979 a long time ago when I wrote it, but is no longer the case.
1981 The emulated memove() is provided for those environments that have neither
1982 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1983 is clearly not the case, as these two bugs were recently reported.
1985 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1986 and Detrail to create the HTML documentation, the .txt form of the man
1987 pages, and it removes trailing spaces from listed files. It also creates
1988 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1989 case, it wraps all the #defines with #ifndefs. This script should be run
1990 before "make dist".
1992 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1993 with Unicode property support.
1995 (a) For a maximizing quantifier, if the two different cases of the
1996 character were of different lengths in their UTF-8 codings (there are
1997 some cases like this - I found 11), and the matching function had to
1998 back up over a mixture of the two cases, it incorrectly assumed they
1999 were both the same length.
2001 (b) When PCRE was configured to use the heap rather than the stack for
2002 recursion during matching, it was not correctly preserving the data for
2003 the other case of a UTF-8 character when checking ahead for a match
2004 while processing a minimizing repeat. If the check also involved
2005 matching a wide character, but failed, corruption could cause an
2006 erroneous result when trying to check for a repeat of the original
2007 character.
2009 13. Some tidying changes to the testing mechanism:
2011 (a) The RunTest script now detects the internal link size and whether there
2012 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2013 values substituted by "configure". (The RunGrepTest script already did
2014 this for UTF-8.) The configure.ac script no longer substitutes the
2015 relevant variables.
2017 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2018 with length and offset values. This means that the output is different
2019 for different internal link sizes. Test 2 is skipped for link sizes
2020 other than 2 because of this, bypassing the problem. Unfortunately,
2021 there was also a test in test 3 (the locale tests) that used /B and
2022 failed for link sizes other than 2. Rather than cut the whole test out,
2023 I have added a new /Z option to pcretest that replaces the length and
2024 offset values with spaces. This is now used to make test 3 independent
2025 of link size. (Test 2 will be tidied up later.)
2027 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2028 segmentation fault instead of returning the appropriate error message.
2030 15. In multiline mode when the newline sequence was set to "any", the pattern
2031 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2032 This doesn't seem right; it now treats the CRLF combination as the line
2033 ending, and so does not match in that case. It's only a pattern such as ^$
2034 that would hit this one: something like ^ABC$ would have failed after \r
2035 and then tried again after \r\n.
2037 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2038 in an attempt to make files that differ only in their line terminators
2039 compare equal. This works on Linux.
2041 17. Under certain error circumstances pcregrep might try to free random memory
2042 as it exited. This is now fixed, thanks to valgrind.
2044 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2045 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2046 was because its rules for how to advance for /g after matching an empty
2047 string at the end of a line did not allow for this case. They now check for
2048 it specially.
2050 20. pcretest is supposed to handle patterns and data of any length, by
2051 extending its buffers when necessary. It was getting this wrong when the
2052 buffer for a data line had to be extended.
2054 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2055 CRLF as a newline sequence.
2057 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2058 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2059 I have nevertheless tidied it up.
2061 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2063 24. Added a man page for pcre-config.
2066 Version 7.0 19-Dec-06
2067 ---------------------
2069 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2070 moving to gcc 4.1.1.
2072 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2073 sys/time.h, which is documented as needed for this function. It doesn't
2074 seem to matter on Linux, but it showed up on some releases of OS X.
2076 3. It seems that there are systems where bytes whose values are greater than
2077 127 match isprint() in the "C" locale. The "C" locale should be the
2078 default when a C program starts up. In most systems, only ASCII printing
2079 characters match isprint(). This difference caused the output from pcretest
2080 to vary, making some of the tests fail. I have changed pcretest so that:
2082 (a) When it is outputting text in the compiled version of a pattern, bytes
2083 other than 32-126 are always shown as hex escapes.
2085 (b) When it is outputting text that is a matched part of a subject string,
2086 it does the same, unless a different locale has been set for the match
2087 (using the /L modifier). In this case, it uses isprint() to decide.
2089 4. Fixed a major bug that caused incorrect computation of the amount of memory
2090 required for a compiled pattern when options that changed within the
2091 pattern affected the logic of the preliminary scan that determines the
2092 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2093 that the computed length was too small. The symptoms of this bug were
2094 either the PCRE error "internal error: code overflow" from pcre_compile(),
2095 or a glibc crash with a message such as "pcretest: free(): invalid next
2096 size (fast)". Examples of patterns that provoked this bug (shown in
2097 pcretest format) are:
2099 /(?-x: )/x
2100 /(?x)(?-x: \s*#\s*)/
2101 /((?i)[\x{c0}])/8
2102 /(?i:[\x{c0}])/8
2104 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2105 is now done differently.
2107 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2108 wrapper classes; (b) implement a new function in the C++ scanner that is
2109 more efficient than the old way of doing things because it avoids levels of
2110 recursion in the regex matching; (c) add a paragraph to the documentation
2111 for the FullMatch() function.
2113 6. The escape sequence \n was being treated as whatever was defined as
2114 "newline". Not only was this contrary to the documentation, which states
2115 that \n is character 10 (hex 0A), but it also went horribly wrong when
2116 "newline" was defined as CRLF. This has been fixed.
2118 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2119 was being set to -1 for the "end of line" case (supposedly a value that no
2120 character can have). Though this value is never used (the check for end of
2121 line is "zero bytes in current character"), it caused compiler complaints.
2122 I've changed it to 0xffffffff.
2124 8. In pcre_version.c, the version string was being built by a sequence of
2125 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2126 string (as it is for production releases) called a macro with an empty
2127 argument. The C standard says the result of this is undefined. The gcc
2128 compiler treats it as an empty string (which was what was wanted) but it is
2129 reported that Visual C gives an error. The source has been hacked around to
2130 avoid this problem.
2132 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2133 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2134 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2135 of them did).
2137 10. Originally, pcretest opened its input and output without "b"; then I was
2138 told that "b" was needed in some environments, so it was added for release
2139 5.0 to both the input and output. (It makes no difference on Unix-like
2140 systems.) Later I was told that it is wrong for the input on Windows. I've
2141 now abstracted the modes into two macros, to make it easier to fiddle with
2142 them, and removed "b" from the input mode under Windows.
2144 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2146 12. Added -help and --help to pcretest as an official way of being reminded
2147 of the options.
2149 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2150 and pcrecpp.cc because they annoy compilers at high warning levels.
2152 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2154 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2155 scripts are not C programs :-) and which was not noticed because it works
2156 on Linux.
2158 16. pcretest is supposed to handle any length of pattern and data line (as one
2159 line or as a continued sequence of lines) by extending its input buffer if
2160 necessary. This feature was broken for very long pattern lines, leading to
2161 a string of junk being passed to pcre_compile() if the pattern was longer
2162 than about 50K.
2164 17. I have done a major re-factoring of the way pcre_compile() computes the
2165 amount of memory needed for a compiled pattern. Previously, there was code
2166 that made a preliminary scan of the pattern in order to do this. That was
2167 OK when PCRE was new, but as the facilities have expanded, it has become
2168 harder and harder to keep it in step with the real compile phase, and there
2169 have been a number of bugs (see for example, 4 above). I have now found a
2170 cunning way of running the real compile function in a "fake" mode that
2171 enables it to compute how much memory it would need, while actually only
2172 ever using a few hundred bytes of working memory and without too many
2173 tests of the mode. This should make future maintenance and development
2174 easier. A side effect of this work is that the limit of 200 on the nesting
2175 depth of parentheses has been removed (though this was never a serious
2176 limitation, I suspect). However, there is a downside: pcre_compile() now
2177 runs more slowly than before (30% or more, depending on the pattern). I
2178 hope this isn't a big issue. There is no effect on runtime performance.
2180 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2181 newline (only possible for the last line of a file) and it was a
2182 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2184 19. Added additional timing features to pcretest. (1) The -tm option now times
2185 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2186 separate command line item, by a number that specifies the number of
2187 repeats to use when timing. The default is 50000; this gives better
2188 precision, but takes uncomfortably long for very large patterns.
2190 20. Extended pcre_study() to be more clever in cases where a branch of a
2191 subpattern has no definite first character. For example, (a*|b*)[cd] would
2192 previously give no result from pcre_study(). Now it recognizes that the
2193 first character must be a, b, c, or d.
2195 21. There was an incorrect error "recursive call could loop indefinitely" if
2196 a subpattern (or the entire pattern) that was being tested for matching an
2197 empty string contained only one non-empty item after a nested subpattern.
2198 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2199 incorrectly, because the \d was being skipped in the check.
2201 22. The pcretest program now has a new pattern option /B and a command line
2202 option -b, which is equivalent to adding /B to every pattern. This causes
2203 it to show the compiled bytecode, without the additional information that
2204 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2205 is the same as /B/I).
2207 23. A new optimization is now able automatically to treat some sequences such
2208 as a*b as a*+b. More specifically, if something simple (such as a character
2209 or a simple class like \d) has an unlimited quantifier, and is followed by
2210 something that cannot possibly match the quantified thing, the quantifier
2211 is automatically "possessified".
2213 24. A recursive reference to a subpattern whose number was greater than 39
2214 went wrong under certain circumstances in UTF-8 mode. This bug could also
2215 have affected the operation of pcre_study().
2217 25. Realized that a little bit of performance could be had by replacing
2218 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2220 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2222 27. Possessive quantifiers such as a++ were previously implemented by turning
2223 them into atomic groups such as ($>a+). Now they have their own opcodes,
2224 which improves performance. This includes the automatically created ones
2225 from 23 above.
2227 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2228 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2229 the first matched character to be a colon. This applied both to named and
2230 numbered groups.
2232 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2234 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2235 building PCRE on Windows easier, so I have included it in the distribution.
2237 31. There is now a check in pcretest against a ridiculously large number being
2238 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2239 loop, the loop is abandoned.
2241 32. Forward references to subpatterns in conditions such as (?(2)...) where
2242 subpattern 2 is defined later cause pcre_compile() to search forwards in
2243 the pattern for the relevant set of parentheses. This search went wrong
2244 when there were unescaped parentheses in a character class, parentheses
2245 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2247 33. "Subroutine" calls and backreferences were previously restricted to
2248 referencing subpatterns earlier in the regex. This restriction has now
2249 been removed.
2251 34. Added a number of extra features that are going to be in Perl 5.10. On the
2252 whole, these are just syntactic alternatives for features that PCRE had
2253 previously implemented using the Python syntax or my own invention. The
2254 other formats are all retained for compatibility.
2256 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2257 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2258 also .NET compatible.
2260 (b) A recursion or subroutine call to a named group can now be defined as
2261 (?&name) as well as (?P>name).
2263 (c) A backreference to a named group can now be defined as \k<name> or
2264 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2265 5.10, are also .NET compatible.
2267 (d) A conditional reference to a named group can now use the syntax
2268 (?(<name>) or (?('name') as well as (?(name).
2270 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2271 groups (named and numbered) that are never evaluated inline, but can be
2272 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2273 is always false. There may be only one alternative in such a group.
2275 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2276 as the simple (?(R). The condition is true only if the most recent
2277 recursion is that of the given number or name. It does not search out
2278 through the entire recursion stack.
2280 (g) The escape \gN or \g{N} has been added, where N is a positive or
2281 negative number, specifying an absolute or relative reference.
2283 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2284 some "unreachable code" warnings.
2286 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2287 things, this adds five new scripts.
2289 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2290 There were also incompatibilities regarding the handling of \Q..\E inside
2291 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2292 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2294 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2295 matches an empty string, and forcibly breaks the loop. There were bugs in
2296 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2297 against aaaa the result was just "a" rather than "aaaa", for example. Two
2298 separate and independent bugs (that affected different cases) have been
2299 fixed.
2301 39. Refactored the code to abolish the use of different opcodes for small
2302 capturing bracket numbers. This is a tidy that I avoided doing when I
2303 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2304 The new approach is not only tidier, it makes it possible to reduce the
2305 memory needed to fix the previous bug (38).
2307 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2308 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2309 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2310 mode.
2312 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2313 report.
2315 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2316 copy construction and assignment in the C++ wrapper.
2318 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2319 couple of bugs that could have given wrong results in the "--newline=crlf"
2320 case.
2322 44. Added a number of casts and did some reorganization of signed/unsigned int
2323 variables following suggestions from Dair Grant. Also renamed the variable
2324 "this" as "item" because it is a C++ keyword.
2326 45. Arranged for dftables to add
2328 #include "pcre_internal.h"
2330 to pcre_chartables.c because without it, gcc 4.x may remove the array
2331 definition from the final binary if PCRE is built into a static library and
2332 dead code stripping is activated.
2334 46. For an unanchored pattern, if a match attempt fails at the start of a
2335 newline sequence, and the newline setting is CRLF or ANY, and the next two
2336 characters are CRLF, advance by two characters instead of one.
2339 Version 6.7 04-Jul-06
2340 ---------------------
2342 1. In order to handle tests when input lines are enormously long, pcretest has
2343 been re-factored so that it automatically extends its buffers when
2344 necessary. The code is crude, but this _is_ just a test program. The
2345 default size has been increased from 32K to 50K.
2347 2. The code in pcre_study() was using the value of the re argument before
2348 testing it for NULL. (Of course, in any sensible call of the function, it
2349 won't be NULL.)
2351 3. The memmove() emulation function in pcre_internal.h, which is used on
2352 systems that lack both memmove() and bcopy() - that is, hardly ever -
2353 was missing a "static" storage class specifier.
2355 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2356 containing an extended class (one that cannot be represented by a bitmap
2357 because it contains high-valued characters or Unicode property items, e.g.
2358 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2359 pattern, but PCRE should not loop if you do not (it no longer does).
2360 [Detail: two cases were found: (a) a repeated subpattern containing an
2361 extended class; (b) a recursive reference to a subpattern that followed a
2362 previous extended class. It wasn't skipping over the extended class
2363 correctly when UTF-8 mode was not set.]
2365 5. A negated single-character class was not being recognized as fixed-length
2366 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2367 compile error "lookbehind assertion is not fixed length".
2369 6. The RunPerlTest auxiliary script was showing an unexpected difference
2370 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2371 write a Perl script that can interpret lines of an input file either as
2372 byte characters or as UTF-8, which is what "perltest" was being required to
2373 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2374 can't do is switch easily at run time between having the "use utf8;" pragma
2375 or not. In the end, I fudged it by using the RunPerlTest script to insert
2376 "use utf8;" explicitly for the UTF-8 tests.
2378 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2379 the end of the subject string, contrary to the documentation and to what
2380 Perl does. This was true of both matching functions. Now it matches only at
2381 the start of the subject and immediately after *internal* newlines.
2383 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2384 a pointer to an int instead of a pointer to an unsigned long int. This
2385 caused problems on 64-bit systems.
2387 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2388 instance of the 'standard' template library not being so standard".
2390 10. There was no check on the number of named subpatterns nor the maximum
2391 length of a subpattern name. The product of these values is used to compute
2392 the size of the memory block for a compiled pattern. By supplying a very
2393 long subpattern name and a large number of named subpatterns, the size
2394 computation could be caused to overflow. This is now prevented by limiting
2395 the length of names to 32 characters, and the number of named subpatterns
2396 to 10,000.
2398 11. Subpatterns that are repeated with specific counts have to be replicated in
2399 the compiled pattern. The size of memory for this was computed from the
2400 length of the subpattern and the repeat count. The latter is limited to
2401 65535, but there was no limit on the former, meaning that integer overflow
2402 could in principle occur. The compiled length of a repeated subpattern is
2403 now limited to 30,000 bytes in order to prevent this.
2405 12. Added the optional facility to have named substrings with the same name.
2407 13. Added the ability to use a named substring as a condition, using the
2408 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2409 are numbers (not recommended). Forward references are permitted.
2411 14. Added forward references in named backreferences (if you see what I mean).
2413 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2414 pattern could run off the end of the subject. For example, the pattern
2415 "(?s)(.{1,5})"8 did this with the subject "ab".
2417 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2418 PCRE_CASELESS was set when matching characters that were quantified with ?
2419 or *.
2421 17. A character class other than a single negated character that had a minimum
2422 but no maximum quantifier - for example [ab]{6,} - was not handled
2423 correctly by pce_dfa_exec(). It would match only one character.
2425 18. A valid (though odd) pattern that looked like a POSIX character
2426 class but used an invalid character after [ (for example [[,abc,]]) caused
2427 pcre_compile() to give the error "Failed: internal error: code overflow" or
2428 in some cases to crash with a glibc free() error. This could even happen if
2429 the pattern terminated after [[ but there just happened to be a sequence of
2430 letters, a binary zero, and a closing ] in the memory that followed.
2432 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2433 over the years. Originally (before any Unicode support), just the bottom 8
2434 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2435 output from "man perlunicode" includes this:
2437 The regular expression compiler produces polymorphic opcodes. That
2438 is, the pattern adapts to the data and automatically switches to
2439 the Unicode character scheme when presented with Unicode data--or
2440 instead uses a traditional byte scheme when presented with byte
2441 data.
2443 Sadly, a wide octal escape does not cause a switch, and in a string with
2444 no other multibyte characters, these octal escapes are treated as before.
2445 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2446 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2447 Unicode string.
2449 I have not perpetrated such confusion in PCRE. Up till now, it took just
2450 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2451 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2452 translate to the appropriate multibyte character.
2454 29. Applied some refactoring to reduce the number of warnings from Microsoft
2455 and Borland compilers. This has included removing the fudge introduced
2456 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2457 a warning about an unused variable.
2459 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2460 characters since release 4.0, because Perl (from release 5.004) does not.
2461 [Or at least, is documented not to: some releases seem to be in conflict
2462 with the documentation.] However, when a pattern was studied with
2463 pcre_study() and all its branches started with \s, PCRE still included VT
2464 as a possible starting character. Of course, this did no harm; it just
2465 caused an unnecessary match attempt.
2467 22. Removed a now-redundant internal flag bit that recorded the fact that case
2468 dependency changed within the pattern. This was once needed for "required
2469 byte" processing, but is no longer used. This recovers a now-scarce options
2470 bit. Also moved the least significant internal flag bit to the most-
2471 significant bit of the word, which was not previously used (hangover from
2472 the days when it was an int rather than a uint) to free up another bit for
2473 the future.
2475 23. Added support for CRLF line endings as well as CR and LF. As well as the
2476 default being selectable at build time, it can now be changed at runtime
2477 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2478 specify that it is scanning data with non-default line endings.
2480 24. Changed the definition of CXXLINK to make it agree with the definition of
2481 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2483 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2484 recursions. This makes a big different to stack usage for some patterns.
2486 26. If a subpattern containing a named recursion or subroutine reference such
2487 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2488 the space required for the compiled pattern went wrong and gave too small a
2489 value. Depending on the environment, this could lead to "Failed: internal
2490 error: code overflow at offset 49" or "glibc detected double free or
2491 corruption" errors.
2493 27. Applied patches from Google (a) to support the new newline modes and (b) to
2494 advance over multibyte UTF-8 characters in GlobalReplace.
2496 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2497 difference for some implementation of PCRE in some Windows version.
2499 29. Added some extra testing facilities to pcretest:
2501 \q<number> in a data line sets the "match limit" value
2502 \Q<number> in a data line sets the "match recursion limt" value
2503 -S <number> sets the stack size, where <number> is in megabytes
2505 The -S option isn't available for Windows.
2508 Version 6.6 06-Feb-06
2509 ---------------------
2511 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2512 in pcreposix.h. I have copied the definition from pcre.h.
2514 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2515 because pcre.h is no longer a built file.
2517 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2518 not normally included in the compiled code.
2521 Version 6.5 01-Feb-06
2522 ---------------------
2524 1. When using the partial match feature with pcre_dfa_exec(), it was not
2525 anchoring the second and subsequent partial matches at the new starting
2526 point. This could lead to incorrect results. For example, with the pattern
2527 /1234/, partially matching against "123" and then "a4" gave a match.
2529 2. Changes to pcregrep:
2531 (a) All non-match returns from pcre_exec() were being treated as failures
2532 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2533 error message is output. Some extra information is given for the
2535 probably the only errors that are likely to be caused by users (by
2536 specifying a regex that has nested indefinite repeats, for instance).
2537 If there are more than 20 of these errors, pcregrep is abandoned.
2539 (b) A binary zero was treated as data while matching, but terminated the
2540 output line if it was written out. This has been fixed: binary zeroes
2541 are now no different to any other data bytes.
2543 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2544 used to set a locale for matching. The --locale=xxxx long option has
2545 been added (no short equivalent) to specify a locale explicitly on the
2546 pcregrep command, overriding the environment variables.
2548 (d) When -B was used with -n, some line numbers in the output were one less
2549 than they should have been.
2551 (e) Added the -o (--only-matching) option.
2553 (f) If -A or -C was used with -c (count only), some lines of context were
2554 accidentally printed for the final match.
2556 (g) Added the -H (--with-filename) option.
2558 (h) The combination of options -rh failed to suppress file names for files
2559 that were found from directory arguments.
2561 (i) Added the -D (--devices) and -d (--directories) options.
2563 (j) Added the -F (--fixed-strings) option.
2565 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2567 (l) Added the --colo(u)r option.
2569 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2570 is not present by default.
2572 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2573 items such as (?R) or (?1), when the recursion could match a number of
2574 alternatives. If it matched one of the alternatives, but subsequently,
2575 outside the recursion, there was a failure, the code tried to back up into
2576 the recursion. However, because of the way PCRE is implemented, this is not
2577 possible, and the result was an incorrect result from the match.
2579 In order to prevent this happening, the specification of recursion has
2580 been changed so that all such subpatterns are automatically treated as
2581 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2583 4. I had overlooked the fact that, in some locales, there are characters for
2584 which isalpha() is true but neither isupper() nor islower() are true. In
2585 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2586 and ordfeminine) are like this. This affected the treatment of \w and \W
2587 when they appeared in character classes, but not when they appeared outside
2588 a character class. The bit map for "word" characters is now created
2589 separately from the results of isalnum() instead of just taking it from the
2590 upper, lower, and digit maps. (Plus the underscore character, of course.)
2592 5. The above bug also affected the handling of POSIX character classes such as
2593 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2594 permanent tables. Instead, the bit maps for such a class were previously
2595 created as the appropriate unions of the upper, lower, and digit bitmaps.
2596 Now they are created by subtraction from the [[:word:]] class, which has
2597 its own bitmap.
2599 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2600 It is created by subtracting the vertical space characters (\x09, \x0a,
2601 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2602 subtraction was done in the overall bitmap for a character class, meaning
2603 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2604 be recognized. This bug has been fixed.
2606 7. Patches from the folks at Google:
2608 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2609 real life, but is still worth protecting against".
2611 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2612 regular expressions".
2614 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2615 have it.
2617 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2618 "configure" and the latter not, in order to fix a problem somebody had
2619 with compiling the Arg class on HP-UX.
2621 (e) Improve the error-handling of the C++ wrapper a little bit.
2623 (f) New tests for checking recursion limiting.
2625 8. The pcre_memmove() function, which is used only if the environment does not
2626 have a standard memmove() function (and is therefore rarely compiled),
2627 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2628 returning a result (though PCRE never actually uses the result).
2630 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2631 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2632 returned instead of calling malloc() with an overflowing number that would
2633 most likely cause subsequent chaos.
2635 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2637 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2638 with this option is matched, the nmatch and pmatch options of regexec() are
2639 ignored.
2641 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2642 provided in case anyone wants to the the POSIX interface with UTF-8
2643 strings.
2645 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2646 C++ linking (needed for some HP-UX environments).
2648 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2649 (unused parameter) and in the pcre_printint() function (omitted "default"
2650 switch label when the default is to do nothing).
2652 15. Added some code to make it possible, when PCRE is compiled as a C++
2653 library, to replace subject pointers for pcre_exec() with a smart pointer
2654 class, thus making it possible to process discontinuous strings.
2656 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2657 much the same function. They were added by different people who were trying
2658 to make PCRE easy to compile on non-Unix systems. It has been suggested
2659 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2660 for compiling on Windows systems. I have therefore replaced it with
2661 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2662 defaults to "extern" for C or "extern C" for C++, which works fine on
2663 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2664 SCOPE with something explicit in config.h. In addition:
2666 (a) pcreposix.h still had just "extern" instead of either of these macros;
2667 I have replaced it with PCRE_DATA_SCOPE.
2669 (b) Functions such as _pcre_xclass(), which are internal to the library,
2670 but external in the C sense, all had PCRE_EXPORT in their definitions.
2671 This is apparently wrong for the Windows case, so I have removed it.
2672 (It makes no difference on Unix-like systems.)
2674 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2675 of recursive calls to match(). This is different to MATCH_LIMIT because
2676 that limits the total number of calls to match(), not all of which increase
2677 the depth of recursion. Limiting the recursion depth limits the amount of
2678 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2679 when PCRE is compiled, and changed at run time. A patch from Google adds
2680 this functionality to the C++ interface.
2682 18. Changes to the handling of Unicode character properties:
2684 (a) Updated the table to Unicode 4.1.0.
2686 (b) Recognize characters that are not in the table as "Cn" (undefined).
2688 (c) I revised the way the table is implemented to a much improved format
2689 which includes recognition of ranges. It now supports the ranges that
2690 are defined in UnicodeData.txt, and it also amalgamates other
2691 characters into ranges. This has reduced the number of entries in the
2692 table from around 16,000 to around 3,000, thus reducing its size
2693 considerably. I realized I did not need to use a tree structure after
2694 all - a binary chop search is just as efficient. Having reduced the
2695 number of entries, I extended their size from 6 bytes to 8 bytes to
2696 allow for more data.
2698 (d) Added support for Unicode script names via properties such as \p{Han}.
2700 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2701 matching that character.
2703 20. When matching a repeated Unicode property with a minimum greater than zero,
2704 (for example \pL{2,}), PCRE could look past the end of the subject if it
2705 reached it while seeking the minimum number of characters. This could
2706 happen only if some of the characters were more than one byte long, because
2707 there is a check for at least the minimum number of bytes.
2709 21. Refactored the implementation of \p and \P so as to be more general, to
2710 allow for more different types of property in future. This has changed the
2711 compiled form incompatibly. Anybody with saved compiled patterns that use
2712 \p or \P will have to recompile them.
2714 22. Added "Any" and "L&" to the supported property types.
2716 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2717 but give a compile time error if the value is greater than 0xff.
2719 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2720 accidentally not being installed or uninstalled.
2722 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2723 made were to insert the current release number. This seemed silly, because
2724 it made things harder for people building PCRE on systems that don't run
2725 "configure". I have turned pcre.h into a distributed file, no longer built
2726 by "configure", with the version identification directly included. There is
2727 no longer a pcre.h.in file.
2729 However, this change necessitated a change to the pcre-config script as
2730 well. It is built from pcre-config.in, and one of the substitutions was the
2731 release number. I have updated configure.ac so that ./configure now finds
2732 the release number by grepping pcre.h.
2734 26. Added the ability to run the tests under valgrind.
2737 Version 6.4 05-Sep-05
2738 ---------------------
2740 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2741 "--" to be printed when multiple files were scanned, even when none of the
2742 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2743 consider it to be a bug, and have restored the previous behaviour.
2745 2. A couple of code tidies to get rid of compiler warnings.
2747 3. The pcretest program used to cheat by referring to symbols in the library
2748 whose names begin with _pcre_. These are internal symbols that are not
2749 really supposed to be visible externally, and in some environments it is
2750 possible to suppress them. The cheating is now confined to including
2751 certain files from the library's source, which is a bit cleaner.
2753 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2754 file's purpose clearer.
2756 5. Reorganized pcre_ucp_findchar().
2759 Version 6.3 15-Aug-05
2760 ---------------------
2762 1. The file libpcre.pc.in did not have general read permission in the tarball.
2764 2. There were some problems when building without C++ support:
2766 (a) If C++ support was not built, "make install" and "make test" still
2767 tried to test it.
2769 (b) There were problems when the value of CXX was explicitly set. Some
2770 changes have been made to try to fix these, and ...
2772 (c) --disable-cpp can now be used to explicitly disable C++ support.
2774 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2775 backslash in a target when C++ was disabled. This confuses some
2776 versions of "make", apparently. Using an intermediate variable solves
2777 this. (Same for CPP_LOBJ.)
2779 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2780 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2781 necessary on certain architectures.
2783 4. Added a setting of -export-symbols-regex to the link command to remove
2784 those symbols that are exported in the C sense, but actually are local
2785 within the library, and not documented. Their names all begin with
2786 "_pcre_". This is not a perfect job, because (a) we have to except some
2787 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2788 available (and never for static libraries). I have made a note to try to
2789 find a way round (a) in the future.
2792 Version 6.2 01-Aug-05
2793 ---------------------
2795 1. There was no test for integer overflow of quantifier values. A construction
2796 such as {1111111111111111} would give undefined results. What is worse, if
2797 a minimum quantifier for a parenthesized subpattern overflowed and became
2798 negative, the calculation of the memory size went wrong. This could have
2799 led to memory overwriting.
2801 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2803 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2804 operating environments where this matters.
2806 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2807 PCRE options from within the C++ wrapper.
2809 5. Named capturing subpatterns were not being correctly counted when a pattern
2810 was compiled. This caused two problems: (a) If there were more than 100
2811 such subpatterns, the calculation of the memory needed for the whole
2812 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2813 back references of the form \12, where the number was greater than 9, were
2814 not recognized as back references, even though there were sufficient
2815 previous subpatterns.
2817 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2818 versions of gcc, e.g. 2.95.4.
2821 Version 6.1 21-Jun-05
2822 ---------------------
2824 1. There was one reference to the variable "posix" in pcretest.c that was not
2825 surrounded by "#if !defined NOPOSIX".
2827 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2828 the cross-check on the old pcre_info() function, for the benefit of the
2829 cut-down version of PCRE that is currently imported into Exim.
2831 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2832 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2833 patterns that start (?i) but I don't think that matters. The use of (?i) is
2834 just an example; this all applies to the other options as well.
2836 4. Since libtool seems to echo the compile commands it is issuing, the output
2837 from "make" can be reduced a bit by putting "@" in front of each libtool
2838 compile command.
2840 5. Patch from the folks at Google for configure.in to be a bit more thorough
2841 in checking for a suitable C++ installation before trying to compile the
2842 C++ stuff. This should fix a reported problem when a compiler was present,
2843 but no suitable headers.
2845 6. The man pages all had just "PCRE" as their title. I have changed them to
2846 be the relevant file name. I have also arranged that these names are
2847 retained in the file doc/pcre.txt, which is a concatenation in text format
2848 of all the man pages except the little individual ones for each function.
2850 7. The NON-UNIX-USE file had not been updated for the different set of source
2851 files that come with release 6. I also added a few comments about the C++
2852 wrapper.
2855 Version 6.0 07-Jun-05
2856 ---------------------
2858 1. Some minor internal re-organization to help with my DFA experiments.
2860 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2861 didn't matter for the library itself when fully configured, but did matter
2862 when compiling without UCP support, or within Exim, where the ucp files are
2863 not imported.
2865 3. Refactoring of the library code to split up the various functions into
2866 different source modules. The addition of the new DFA matching code (see
2867 below) to a single monolithic source would have made it really too
2868 unwieldy, quite apart from causing all the code to be include in a
2869 statically linked application, when only some functions are used. This is
2870 relevant even without the DFA addition now that patterns can be compiled in
2871 one application and matched in another.
2873 The downside of splitting up is that there have to be some external
2874 functions and data tables that are used internally in different modules of
2875 the library but which are not part of the API. These have all had their
2876 names changed to start with "_pcre_" so that they are unlikely to clash
2877 with other external names.
2879 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2880 a different (DFA) algorithm. Although it is slower than the original
2881 function, it does have some advantages for certain types of matching
2882 problem.
2884 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2885 including restarting after a partial match.
2887 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2888 defined when compiling for Windows was sent to me. I have put it into the
2889 code, though I have no means of testing or verifying it.
2891 7. Added the pcre_refcount() auxiliary function.
2893 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2894 match before or at the first newline in the subject string. In pcretest,
2895 the /f option on a pattern can be used to set this.
2897 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2898 would behave wrongly. This has been present in PCRE since release 4.0.
2900 10. A number of changes to the pcregrep command:
2902 (a) Refactored how -x works; insert ^(...)$ instead of setting
2903 PCRE_ANCHORED and checking the length, in preparation for adding
2904 something similar for -w.
2906 (b) Added the -w (match as a word) option.
2908 (c) Refactored the way lines are read and buffered so as to have more
2909 than one at a time available.
2911 (d) Implemented a pcregrep test script.
2913 (e) Added the -M (multiline match) option. This allows patterns to match
2914 over several lines of the subject. The buffering ensures that at least
2915 8K, or the rest of the document (whichever is the shorter) is available
2916 for matching (and similarly the previous 8K for lookbehind assertions).
2918 (f) Changed the --help output so that it now says
2920 -w, --word-regex(p)
2922 instead of two lines, one with "regex" and the other with "regexp"
2923 because that confused at least one person since the short forms are the
2924 same. (This required a bit of code, as the output is generated
2925 automatically from a table. It wasn't just a text change.)
2927 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2928 option but starts with a hyphen. Could be a pattern or a path name
2929 starting with a hyphen, for instance.
2931 (h) "-" can be given as a file name to represent stdin.
2933 (i) When file names are being printed, "(standard input)" is used for
2934 the standard input, for compatibility with GNU grep. Previously
2935 "<stdin>" was used.
2937 (j) The option --label=xxx can be used to supply a name to be used for
2938 stdin when file names are being printed. There is no short form.
2940 (k) Re-factored the options decoding logic because we are going to add
2941 two more options that take data. Such options can now be given in four
2942 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2944 (l) Added the -A, -B, and -C options for requesting that lines of context
2945 around matches be printed.
2947 (m) Added the -L option to print the names of files that do not contain
2948 any matching lines, that is, the complement of -l.
2950 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2951 continue to scan other files.
2953 (o) The -s option was incorrectly implemented. For compatibility with other
2954 greps, it now suppresses the error message for a non-existent or non-
2955 accessible file (but not the return code). There is a new option called
2956 -q that suppresses the output of matching lines, which was what -s was
2957 previously doing.
2959 (p) Added --include and --exclude options to specify files for inclusion
2960 and exclusion when recursing.
2962 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2963 Hopefully, it now does.
2965 12. Missing cast in pcre_study().
2967 13. Added an "uninstall" target to the makefile.
2969 14. Replaced "extern" in the function prototypes in Makefile.in with
2970 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2971 world, but is set differently for Windows.
2973 15. Added a second compiling function called pcre_compile2(). The only
2974 difference is that it has an extra argument, which is a pointer to an
2975 integer error code. When there is a compile-time failure, this is set
2976 non-zero, in addition to the error test pointer being set to point to an
2977 error message. The new argument may be NULL if no error number is required
2978 (but then you may as well call pcre_compile(), which is now just a
2979 wrapper). This facility is provided because some applications need a
2980 numeric error indication, but it has also enabled me to tidy up the way
2981 compile-time errors are handled in the POSIX wrapper.
2983 16. Added VPATH=.libs to the makefile; this should help when building with one
2984 prefix path and installing with another. (Or so I'm told by someone who
2985 knows more about this stuff than I do.)
2987 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2988 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2989 match everything, including newlines. This is not POSIX-compatible, but
2990 somebody wanted the feature. From pcretest it can be activated by using
2991 both the P and the s flags.
2993 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2995 19. libpcre.pc was being incorrectly installed as executable.
2997 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2998 it now also looks for '\r' so that it will work unmodified on Windows.
3000 21. Added Google's contributed C++ wrapper to the distribution.
3002 22. Added some untidy missing memory free() calls in pcretest, to keep
3003 Electric Fence happy when testing.
3007 Version 5.0 13-Sep-04
3008 ---------------------
3010 1. Internal change: literal characters are no longer packed up into items
3011 containing multiple characters in a single byte-string. Each character
3012 is now matched using a separate opcode. However, there may be more than one
3013 byte in the character in UTF-8 mode.
3015 2. The pcre_callout_block structure has two new fields: pattern_position and
3016 next_item_length. These contain the offset in the pattern to the next match
3017 item, and its length, respectively.
3019 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3020 insertion of callouts before each pattern item. Added the /C option to
3021 pcretest to make use of this.
3023 4. On the advice of a Windows user, the lines
3025 #if defined(_WIN32) || defined(WIN32)
3026 _setmode( _fileno( stdout ), 0x8000 );
3027 #endif /* defined(_WIN32) || defined(WIN32) */
3029 have been added to the source of pcretest. This apparently does useful
3030 magic in relation to line terminators.
3032 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3033 for the benefit of those environments where the "b" makes a difference.
3035 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3036 to know about it. I have put a hack into configure.in that adds in code
3037 to set GCC=yes if CC=icc. This seems to end up at a point in the
3038 generated configure script that is early enough to affect the setting of
3039 compiler options, which is what is needed, but I have no means of testing
3040 whether it really works. (The user who reported this had patched the
3041 generated configure script, which of course I cannot do.)
3043 LATER: After change 22 below (new libtool files), the configure script
3044 seems to know about icc (and also ecc). Therefore, I have commented out
3045 this hack in configure.in.
3047 7. Added support for pkg-config (2 patches were sent in).
3049 8. Negated POSIX character classes that used a combination of internal tables
3050 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3051 [[:^ascii]]. Typically, they would match almost any characters. The other
3052 POSIX classes were not broken in this way.
3054 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3055 to find the match, as PCRE was deluded into thinking that the match had to
3056 start at the start point or following a newline. The same bug applied to
3057 patterns with negative forward assertions or any backward assertions
3058 preceding ".*" at the start, unless the pattern required a fixed first
3059 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3061 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3062 starting at the last subject character, bytes beyond the end of the subject
3063 string were read.
3065 11. Renamed the variable "class" as "classbits" to make life easier for C++
3066 users. (Previously there was a macro definition, but it apparently wasn't
3067 enough.)
3069 12. Added the new field "tables" to the extra data so that tables can be passed
3070 in at exec time, or the internal tables can be re-selected. This allows
3071 a compiled regex to be saved and re-used at a later time by a different
3072 program that might have everything at different addresses.
3074 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3075 -R library as well as a -L library.
3077 14. The debugging options of pcretest (-d on the command line or D on a
3078 pattern) showed incorrect output for anything following an extended class
3079 that contained multibyte characters and which was followed by a quantifier.
3081 15. Added optional support for general category Unicode character properties
3082 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3083 support. It adds about 90K to the size of the library. The meanings of the
3084 inbuilt class escapes such as \d and \s have NOT been changed.
3086 16. Updated pcredemo.c to include calls to free() to release the memory for the
3087 compiled pattern.
3089 17. The generated file chartables.c was being created in the source directory
3090 instead of in the building directory. This caused the build to fail if the
3091 source directory was different from the building directory, and was
3092 read-only.
3094 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3095 file. No doubt somebody will tell me if they don't make sense... Also added
3096 Dan Mooney's comments about building on OpenVMS.
3098 19. Added support for partial matching via the PCRE_PARTIAL option for
3099 pcre_exec() and the \P data escape in pcretest.
3101 20. Extended pcretest with 3 new pattern features:
3103 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3104 write the compiled pattern to the file whose name is "rest-of-line".
3105 This is a straight binary dump of the data, with the saved pointer to
3106 the character tables forced to be NULL. The study data, if any, is
3107 written too. After writing, pcretest reads a new pattern.
3109 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3110 compiled pattern from the given file. There must not be any
3111 occurrences of "<" in the file name (pretty unlikely); if there are,
3112 pcretest will instead treat the initial "<" as a pattern delimiter.
3113 After reading in the pattern, pcretest goes on to read data lines as
3114 usual.
3116 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3117 and 16-bit fields in a compiled pattern, to simulate a pattern that
3118 was compiled on a host of opposite endianness.
3120 21. The pcre-exec() function can now cope with patterns that were compiled on
3121 hosts of opposite endianness, with this restriction:
3123 As for any compiled expression that is saved and used later, the tables
3124 pointer field cannot be preserved; the extra_data field in the arguments
3125 to pcre_exec() should be used to pass in a tables address if a value
3126 other than the default internal tables were used at compile time.
3128 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3129 now diagnosed as an error. Previously, most of the time, a negative number
3130 would have been treated as zero, but if in addition "ovector" was passed as
3131 NULL, a crash could occur.
3133 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3134 new versions from the libtool 1.5 distribution (the last one is a copy of
3135 a file called libtool.m4). This seems to have fixed the need to patch
3136 "configure" to support Darwin 1.3 (which I used to do). However, I still
3137 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3138 workstation).
3140 24. Changed the PCRE licence to be the more standard "BSD" licence.
3143 Version 4.5 01-Dec-03
3144 ---------------------
3146 1. There has been some re-arrangement of the code for the match() function so
3147 that it can be compiled in a version that does not call itself recursively.
3148 Instead, it keeps those local variables that need separate instances for
3149 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3150 needs to "recurse". Keeping track of where control must go is done by means
3151 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3152 hide most of the details from the main code, and operates only if
3153 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3154 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3155 operating.
3157 To make it easier for callers to provide specially tailored get/free
3158 functions for this usage, two new functions, pcre_stack_malloc, and
3159 pcre_stack_free, are used. They are always called in strict stacking order,
3160 and the size of block requested is always the same.
3162 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3163 PCRE has been compiled to use the stack or the heap for recursion. The
3164 -C option of pcretest uses this to show which version is compiled.
3166 A new data escape \S, is added to pcretest; it causes the amounts of store
3167 obtained and freed by both kinds of malloc/free at match time to be added
3168 to the output.
3170 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3171 what's available on my current Linux desktop machine.
3173 3. When matching a UTF-8 string, the test for a valid string at the start has
3174 been extended. If start_offset is not zero, PCRE now checks that it points
3175 to a byte that is the start of a UTF-8 character. If not, it returns
3176 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3177 this is necessary because there may be backward assertions in the pattern.
3178 When matching the same subject several times, it may save resources to use
3179 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3181 4. The code for checking the validity of UTF-8 strings has been tightened so
3182 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3183 containing "overlong sequences".
3185 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3186 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3187 should have been "&", but it just so happened that all the cases this let
3188 through by mistake were picked up later in the function.
3190 6. I had used a variable called "isblank" - this is a C99 function, causing
3191 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3193 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3194 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3195 so the version is shown; (c) stop "make test" echoing "./RunTest".
3197 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3199 9. The prototype for memmove() for systems that don't have it was using
3200 size_t, but the inclusion of the header that defines size_t was later. I've
3201 moved the #includes for the C headers earlier to avoid this.
3203 10. Added some adjustments to the code to make it easier to compiler on certain
3204 special systems:
3206 (a) Some "const" qualifiers were missing.
3207 (b) Added the macro EXPORT before all exported functions; by default this
3208 is defined to be empty.
3209 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3210 that it reads its output file name as an argument instead of writing
3211 to the standard output and assuming this can be redirected.
3213 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3214 class containing characters with values greater than 255, PCRE compilation
3215 went into a loop.
3217 12. A recursive reference to a subpattern that was within another subpattern
3218 that had a minimum quantifier of zero caused PCRE to crash. For example,
3219 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3220 recursion. If the recursively-called subpattern itself had a zero repeat,
3221 that was OK.
3223 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3224 buffer into which it was copied (for escape processing) was still set at
3225 1024, so long lines caused crashes.
3227 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3228 "internal error: code overflow...". This applied to any character class
3229 that was followed by a possessive quantifier.
3231 15. Modified the Makefile to add libpcre.la as a prerequisite for
3232 libpcreposix.la because I was told this is needed for a parallel build to
3233 work.
3235 16. If a pattern that contained .* following optional items at the start was
3236 studied, the wrong optimizing data was generated, leading to matching
3237 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3238 matching string must start with a or b or c. The correct conclusion for
3239 this pattern is that a match can start with any character.
3242 Version 4.4 13-Aug-03
3243 ---------------------
3245 1. In UTF-8 mode, a character class containing characters with values between
3246 127 and 255 was not handled correctly if the compiled pattern was studied.
3247 In fixing this, I have also improved the studying algorithm for such
3248 classes (slightly).
3250 2. Three internal functions had redundant arguments passed to them. Removal
3251 might give a very teeny performance improvement.
3253 3. Documentation bug: the value of the capture_top field in a callout is *one
3254 more than* the number of the hightest numbered captured substring.
3256 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3257 in incorrectly linking with a previously installed version. They now link
3258 explicitly with libpcre.la.
3260 5. configure.in no longer needs to recognize Cygwin specially.
3262 6. A problem in pcre.in for Windows platforms is fixed.
3264 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3265 pcretest, it used to include the size of the study block as part of its
3266 output. Unfortunately, the structure contains a field that has a different
3267 size on different hardware architectures. This meant that the tests that
3268 showed this size failed. As the block is currently always of a fixed size,
3269 this information isn't actually particularly useful in pcretest output, so
3270 I have just removed it.
3272 8. Three pre-processor statements accidentally did not start in column 1.
3273 Sadly, there are *still* compilers around that complain, even though
3274 standard C has not required this for well over a decade. Sigh.
3276 9. In pcretest, the code for checking callouts passed small integers in the
3277 callout_data field, which is a void * field. However, some picky compilers
3278 complained about the casts involved for this on 64-bit systems. Now
3279 pcretest passes the address of the small integer instead, which should get
3280 rid of the warnings.
3282 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3283 both compile and run time, and gives an error if an invalid UTF-8 sequence
3284 is found. There is a option for disabling this check in cases where the
3285 string is known to be correct and/or the maximum performance is wanted.
3287 11. In response to a bug report, I changed one line in Makefile.in from
3289 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3290 to
3291 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3293 to look similar to other lines, but I have no way of telling whether this
3294 is the right thing to do, as I do not use Windows. No doubt I'll get told
3295 if it's wrong...
3298 Version 4.3 21-May-03
3299 ---------------------
3301 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3302 Makefile.
3304 2. Some refactoring to improve the quality of the code:
3306 (i) The utf8_table... variables are now declared "const".
3308 (ii) The code for \cx, which used the "case flipping" table to upper case
3309 lower case letters, now just substracts 32. This is ASCII-specific,
3310 but the whole concept of \cx is ASCII-specific, so it seems
3311 reasonable.
3313 (iii) PCRE was using its character types table to recognize decimal and
3314 hexadecimal digits in the pattern. This is silly, because it handles
3315 only 0-9, a-f, and A-F, but the character types table is locale-
3316 specific, which means strange things might happen. A private
3317 table is now used for this - though it costs 256 bytes, a table is
3318 much faster than multiple explicit tests. Of course, the standard
3319 character types table is still used for matching digits in subject
3320 strings against \d.
3322 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3323 ending in _t are). So I've renamed it as ESC_tee.
3325 3. The first argument for regexec() in the POSIX wrapper should have been
3326 defined as "const".
3328 4. Changed pcretest to use malloc() for its buffers so that they can be
3329 Electric Fenced for debugging.
3331 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3332 to read one or more bytes before the start of the subject string. Often this
3333 had no effect on PCRE's behaviour, but in some circumstances it could
3334 provoke a segmentation fault.
3336 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3337 to try to read one or more bytes before the start of the subject string.
3339 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3340 UTF-8 support could misbehave in various ways if the subject string
3341 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3342 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3343 back over UTF-8 characters.)
3346 Version 4.2 14-Apr-03
3347 ---------------------
3349 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3351 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3352 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3353 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3354 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3355 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3357 Note: automatic setting of the BUILD variables is not yet working
3358 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3359 compile-time but not at link-time
3360 [LINK]: use for linking executables only
3361 make different versions for Windows and non-Windows
3362 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3363 libraries
3364 [LINK_FOR_BUILD]: new variable
3365 [OBJEXT]: use throughout
3366 [EXEEXT]: use throughout
3367 <winshared>: new target
3368 <wininstall>: new target
3369 <dftables.o>: use native compiler
3370 <dftables>: use native linker
3371 <install>: handle Windows platform correctly
3372 <clean>: ditto
3373 <check>: ditto
3374 copy DLL to top builddir before testing
3376 As part of these changes, -no-undefined was removed again. This was reported
3377 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3378 in any case.
3380 3. Some tidies to get rid of compiler warnings:
3382 . In the match_data structure, match_limit was an unsigned long int, whereas
3383 match_call_count was an int. I've made them both unsigned long ints.
3385 . In pcretest the fact that a const uschar * doesn't automatically cast to
3386 a void * provoked a warning.
3388 . Turning on some more compiler warnings threw up some "shadow" variables
3389 and a few more missing casts.
3391 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3392 option, a class that contained a single character with a value between 128
3393 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3395 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3396 option, a class that contained several characters, but with at least one
3397 whose value was between 128 and 255 caused PCRE to crash.
3400 Version 4.1 12-Mar-03
3401 ---------------------
3403 1. Compiling with gcc -pedantic found a couple of places where casts were
3404 needed, and a string in dftables.c that was longer than standard compilers are
3405 required to support.
3407 2. Compiling with Sun's compiler found a few more places where the code could
3408 be tidied up in order to avoid warnings.
3410 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3411 first of these names is deprecated in the latest Autoconf in favour of the name
3412 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3413 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3414 analogy I have changed it to CFLAGS_FOR_BUILD.
3416 4. Added -no-undefined to the linking command in the Makefile, because this is
3417 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3418 linking step for the pcreposix library.
3420 5. PCRE was failing to diagnose the case of two named groups with the same
3421 name.
3423 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3424 literal character that is needed in the subject for a match, and scans along to
3425 ensure that it is present before embarking on the full matching process. This
3426 saves time in cases of nested unlimited repeats that are never going to match.
3427 Problem: the scan can take a lot of time if the subject is very long (e.g.
3428 megabytes), thus penalizing straightforward matches. It is now done only if the
3429 amount of subject to be scanned is less than 1000 bytes.
3431 7. A lesser problem with the same optimization is that it was recording the
3432 first character of an anchored pattern as "needed", thus provoking a search
3433 right along the subject, even when the first match of the pattern was going to
3434 fail. The "needed" character is now not set for anchored patterns, unless it
3435 follows something in the pattern that is of non-fixed length. Thus, it still
3436 fulfils its original purpose of finding quick non-matches in cases of nested
3437 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3440 Version 4.0 17-Feb-03
3441 ---------------------
3443 1. If a comment in an extended regex that started immediately after a meta-item
3444 extended to the end of string, PCRE compiled incorrect data. This could lead to
3445 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3447 2. Moved to autoconf 2.53 and libtool 1.4.2.
3449 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3450 the special perltest8 script is no longer needed - all the tests can be run
3451 from a single perltest script.
3453 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3454 by \s. It has now been removed in PCRE. This means it isn't recognized as
3455 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3456 class [:space:] *does* include VT, thereby creating a mess.
3458 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3459 space and tab.
3461 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3462 its new features into the main test script, reducing the number of scripts.
3464 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3465 were backward compatible, and made the (?i) apply to the whole pattern, as if
3466 /i were given. Now it behaves more logically, and applies the option setting
3467 only to what follows. PCRE has been changed to follow suit. However, if it
3468 finds options settings right at the start of the pattern, it extracts them into
3469 the global options, as before. Thus, they show up in the info data.
3471 8. Added support for the \Q...\E escape sequence. Characters in between are
3472 treated as literals. This is slightly different from Perl in that $ and @ are
3473 also handled as literals inside the quotes. In Perl, they will cause variable
3474 interpolation. Note the following examples:
3476 Pattern PCRE matches Perl matches
3478 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3479 \Qabc\$xyz\E abc\$xyz abc\$xyz
3480 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3482 For compatibility with Perl, \Q...\E sequences are recognized inside character
3483 classes as well as outside them.
3485 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3486 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3487 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3488 signed/unsigned warnings.
3490 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3491 option for pcretest, so I've replaced it by a simple function that does just
3492 that job.
3494 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3495 "pcregrep -".
3497 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3498 Java package. This provides some syntactic sugar for simple cases of what my
3499 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3500 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3501 item, you can use this simplified notation. Note that only makes sense with
3502 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3503 greediness, whatever the setting of the PCRE_UNGREEDY option.
3505 13. A change of greediness default within a pattern was not taking effect at
3506 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3507 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3508 was abstracted outside.
3510 14. PCRE now supports the \G assertion. It is true when the current matching
3511 position is at the start point of the match. This differs from \A when the
3512 starting offset is non-zero. Used with the /g option of pcretest (or similar
3513 code), it works in the same way as it does for Perl's /g option. If all
3514 alternatives of a regex begin with \G, the expression is anchored to the start
3515 match position, and the "anchored" flag is set in the compiled expression.
3517 15. Some bugs concerning the handling of certain option changes within patterns
3518 have been fixed. These applied to options other than (?ims). For example,
3519 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3520 been the other way round. Some of this was related to change 7 above.
3522 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3523 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3524 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3525 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3527 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3528 mode. Unlike ".", it always matches newline, whatever the setting of
3529 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3530 assertions. Perl allows it, but it doesn't (in general) work because it can't
3531 calculate the length of the lookbehind. At least, that's the case for Perl
3532 5.8.0 - I've been told they are going to document that it doesn't work in
3533 future.
3535 18. Added an error diagnosis for escapes that PCRE does not support: these are
3536 \L, \l, \N, \P, \p, \U, \u, and \X.
3538 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3539 reading past the end of the pattern in cases such as /[abcd/.
3541 20. PCRE was getting more memory than necessary for patterns with classes that
3542 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3544 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3545 compiling PCRE for use with Virtual Pascal.
3547 22. Small fix to the Makefile to make it work properly if the build is done
3548 outside the source tree.
3550 23. Added a new extension: a condition to go with recursion. If a conditional
3551 subpattern starts with (?(R) the "true" branch is used if recursion has
3552 happened, whereas the "false" branch is used only at the top level.
3554 24. When there was a very long string of literal characters (over 255 bytes
3555 without UTF support, over 250 bytes with UTF support), the computation of how
3556 much memory was required could be incorrect, leading to segfaults or other
3557 strange effects.
3559 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3560 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3561 there was a subsequent back reference to those brackets. This meant that, for
3562 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3563 possible to check for precisely this case. All we can do is abandon the
3564 optimization if .* occurs inside capturing brackets when there are any back
3565 references whatsoever. (See below for a better fix that came later.)
3567 26. The handling of the optimization for finding the first character of a
3568 non-anchored pattern, and for finding a character that is required later in the
3569 match were failing in some cases. This didn't break the matching; it just
3570 failed to optimize when it could. The way this is done has been re-implemented.
3572 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3574 28. Added a new feature that provides some of the functionality that Perl
3575 provides with (?{...}). The facility is termed a "callout". The way it is done
3576 in PCRE is for the caller to provide an optional function, by setting
3577 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3578 global variable. By default it is unset, which disables all calling out. To get
3579 the function called, the regex must include (?C) at appropriate points. This
3580 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3581 This provides a means of identifying different callout points. When PCRE
3582 reaches such a point in the regex, if pcre_callout has been set, the external
3583 function is called. It is provided with data in a structure called
3584 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3585 matching continues; if it returns a non-zero value, the match at the current
3586 point fails. However, backtracking will occur if possible. [This was changed
3587 later and other features added - see item 49 below.]
3589 29. pcretest is upgraded to test the callout functionality. It provides a
3590 callout function that displays information. By default, it shows the start of
3591 the match and the current position in the text. There are some new data escapes
3592 to vary what happens:
3594 \C+ in addition, show current contents of captured substrings
3595 \C- do not supply a callout function
3596 \C!n return 1 when callout number n is reached
3597 \C!n!m return 1 when callout number n is reached for the mth time
3599 30. If pcregrep was called with the -l option and just a single file name, it
3600 output "<stdin>" if a match was found, instead of the file name.
3602 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3603 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3604 pcre_exec(). This saves a malloc/free per call. The default value of
3605 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3606 when configuring.
3608 32. The default maximum size of a compiled pattern is 64K. There have been a
3609 few cases of people hitting this limit. The code now uses macros to handle the
3610 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3611 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3612 configuring. Tests 2 and 5 work only with 2-byte links because they output
3613 debugging information about compiled patterns.
3615 33. Internal code re-arrangements:
3617 (a) Moved the debugging function for printing out a compiled regex into
3618 its own source file (printint.c) and used #include to pull it into
3619 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3620 separate copies.
3622 (b) Defined the list of op-code names for debugging as a macro in
3623 internal.h so that it is next to the definition of the opcodes.
3625 (c) Defined a table of op-code lengths for simpler skipping along compiled
3626 code. This is again a macro in internal.h so that it is next to the
3627 definition of the opcodes.
3629 34. Added support for recursive calls to individual subpatterns, along the
3630 lines of Robin Houston's patch (but implemented somewhat differently).
3632 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3633 allow it to read and process whole directories in Win32. This code was
3634 contributed by Lionel Fourquaux; it has not been tested by me.
3636 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3637 used to name a group. Names consist of alphanumerics and underscores, and must
3638 be unique. Back references use the syntax (?P=name) and recursive calls use
3639 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3640 numbers. The function pcre_fullinfo() can be used after compilation to extract
3641 a name/number map. There are three relevant calls:
3643 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3644 PCRE_INFO_NAMECOUNT yields the number of entries
3645 PCRE_INFO_NAMETABLE yields a pointer to the map.
3647 The map is a vector of fixed-size entries. The size of each entry depends on
3648 the length of the longest name used. The first two bytes of each entry are the
3649 group number, most significant byte first. There follows the corresponding
3650 name, zero terminated. The names are in alphabetical order.
3652 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3653 case instead of 255. Making it the same both with and without UTF-8 support
3654 means that the same test output works with both.
3656 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3657 calling malloc() with a zero argument.
3659 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3660 optimization. I've improved things by keeping a bitmap of backreferences with
3661 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3662 fact referenced, the optimization can be applied. It is unlikely that a
3663 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3664 the match to follow \n) will appear inside brackets with a number greater than
3665 31, but if it does, any back reference > 31 suppresses the optimization.
3667 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3668 of disabling numbered capturing parentheses. Any opening parenthesis that is
3669 not followed by ? behaves as if it were followed by ?: but named parentheses
3670 can still be used for capturing (and they will acquire numbers in the usual
3671 way).
3673 41. Redesigned the return codes from the match() function into yes/no/error so
3674 that errors can be passed back from deep inside the nested calls. A malloc
3675 failure while inside a recursive subpattern call now causes the
3676 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3678 42. It is now possible to set a limit on the number of times the match()
3679 function is called in a call to pcre_exec(). This facility makes it possible to
3680 limit the amount of recursion and backtracking, though not in a directly
3681 obvious way, because the match() function is used in a number of different
3682 circumstances. The count starts from zero for each position in the subject
3683 string (for non-anchored patterns). The default limit is, for compatibility, a
3684 large number, namely 10 000 000. You can change this in two ways:
3686 (a) When configuring PCRE before making, you can use --with-match-limit=n
3687 to set a default value for the compiled library.
3689 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3690 a different value is set. See 45 below.
3692 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3694 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3695 of things that can be changed at compile time. The first argument specifies
3696 what is wanted and the second points to where the information is to be placed.
3697 The current list of available information is:
3701 The output is an integer that is set to one if UTF-8 support is available;
3702 otherwise it is set to zero.
3706 The output is an integer that it set to the value of the code that is used for
3707 newline. It is either LF (10) or CR (13).
3711 The output is an integer that contains the number of bytes used for internal
3712 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3716 The output is an integer that contains the threshold above which the POSIX
3717 interface uses malloc() for output vectors. See item 31 above.
3721 The output is an unsigned integer that contains the default limit of the number
3722 of match() calls in a pcre_exec() execution. See 42 above.
3724 44. pcretest has been upgraded by the addition of the -C option. This causes it
3725 to extract all the available output from the new pcre_config() function, and to
3726 output it. The program then exits immediately.
3728 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3729 order to support additional features. One way would have been to define
3730 pcre_exec2() (for example) with extra arguments, but this would not have been
3731 extensible, and would also have required all calls to the original function to
3732 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3733 is used for passing in "extra" data from pcre_study().
3735 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3736 contains the following fields:
3738 flags a bitmap indicating which of the following fields are set
3739 study_data opaque data from pcre_study()
3740 match_limit a way of specifying a limit on match() calls for a specific
3741 call to pcre_exec()
3742 callout_data data for callouts (see 49 below)
3744 The flag bits are also defined in pcre.h, and are
3750 The pcre_study() function now returns one of these new pcre_extra blocks, with
3751 the actual study data pointed to by the study_data field, and the
3752 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3753 before. That is, this change is entirely upwards-compatible and requires no
3754 change to existing code.
3756 If you want to pass in additional data to pcre_exec(), you can either place it
3757 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3758 block.
3760 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3761 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3762 times with different match limits, until it finds the minimum value needed for
3763 pcre_exec() to complete. The value is then output. This can be instructive; for
3764 most simple matches the number is quite small, but for pathological cases it
3765 gets very large very quickly.
3767 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3768 returns the size of the data block pointed to by the study_data field in a
3769 pcre_extra block, that is, the value that was passed as the argument to
3770 pcre_malloc() when PCRE was getting memory in which to place the information
3771 created by pcre_study(). The fourth argument should point to a size_t variable.
3772 pcretest has been extended so that this information is shown after a successful
3773 pcre_study() call when information about the compiled regex is being displayed.
3775 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3776 because what follows is always an absolute path. (Later: it turns out that this
3777 is more than cosmetic for MinGW, because it doesn't like empty path
3778 components.)
3780 49. Some changes have been made to the callout feature (see 28 above):
3782 (i) A callout function now has three choices for what it returns:
3784 0 => success, carry on matching
3785 > 0 => failure at this point, but backtrack if possible
3786 < 0 => serious error, return this value from pcre_exec()
3788 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3789 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3790 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3791 use by callout functions. It will never be used by PCRE itself.
3793 (ii) The pcre_extra structure (see 45 above) has a void * field called
3794 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3795 pcre_callout_block structure has a field of the same name. The contents of
3796 the field passed in the pcre_extra structure are passed to the callout
3797 function in the corresponding field in the callout block. This makes it
3798 easier to use the same callout-containing regex from multiple threads. For
3799 testing, the pcretest program has a new data escape
3801 \C*n pass the number n (may be negative) as callout_data
3803 If the callout function in pcretest receives a non-zero value as
3804 callout_data, it returns that value.
3806 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3807 there were some redundant $(CFLAGS) in commands that are now specified as
3808 $(LINK), which already includes $(CFLAGS).
3810 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3811 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3812 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3813 one-byte characters throughout. Note that case-insensitive matching applies
3814 only to characters whose values are less than 256. PCRE doesn't support the
3815 notion of cases for higher-valued characters.
3817 (i) A character class whose characters are all within 0-255 is handled as
3818 a bit map, and the map is inverted for negative classes. Previously, a
3819 character > 255 always failed to match such a class; however it should
3820 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3822 (ii) A negated character class with a single character < 255 is coded as
3823 "not this character" (OP_NOT). This wasn't working properly when the test
3824 character was multibyte, either singly or repeated.
3826 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3827 mode, for example: \x{100}{2,3}.
3829 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3830 singly or repeated) now correctly test multibyte characters. However,
3831 PCRE doesn't recognize any characters with values greater than 255 as
3832 digits, spaces, or word characters. Such characters always match \D, \S,
3833 and \W, and never match \d, \s, or \w.
3835 (v) Classes may now contain characters and character ranges with values
3836 greater than 255. For example: [ab\x{100}-\x{400}].
3838 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3839 PCRE in UTF-8 mode.
3841 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3842 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3843 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3844 value.)
3846 53. The single man page has become too large. I have therefore split it up into
3847 a number of separate man pages. These also give rise to individual HTML pages;
3848 these are now put in a separate directory, and there is an index.html page that
3849 lists them all. Some hyperlinking between the pages has been installed.
3851 54. Added convenience functions for handling named capturing parentheses.
3853 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3854 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3855 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3856 are faulted.
3858 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3859 calling configure. These values are used when compiling the dftables.c program
3860 which is run to generate the source of the default character tables. They
3861 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3862 you will need to set these values.
3864 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3867 Version 3.9 02-Jan-02
3868 ---------------------
3870 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3872 2. If --disable-static was given, the building process failed when trying to
3873 build pcretest and pcregrep. (For some reason it was using libtool to compile
3874 them, which is not right, as they aren't part of the library.)
3877 Version 3.8 18-Dec-01
3878 ---------------------
3880 1. The experimental UTF-8 code was completely screwed up. It was packing the
3881 bytes in the wrong order. How dumb can you get?
3884 Version 3.7 29-Oct-01
3885 ---------------------
3887 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3888 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3889 this didn't happen under Solaris 8, where I normally test things.
3891 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3892 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3893 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3894 this fix an hour or so after the initial 3.7 release.)
3897 Version 3.6 23-Oct-01
3898 ---------------------
3900 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3901 offsets passed as NULL with zero offset count.
3903 2. The config.guess and config.sub files had not been updated when I moved to
3904 the latest autoconf.
3907 Version 3.5 15-Aug-01
3908 ---------------------
3910 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3911 had been forgotten.
3913 2. By using declared but undefined structures, we can avoid using "void"
3914 definitions in pcre.h while keeping the internal definitions of the structures
3915 private.
3917 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3918 user point of view, this means that both static and shared libraries are built
3919 by default, but this can be individually controlled. More of the work of
3920 handling this static/shared cases is now inside libtool instead of PCRE's make
3921 file.
3923 4. The pcretest utility is now installed along with pcregrep because it is
3924 useful for users (to test regexs) and by doing this, it automatically gets
3925 relinked by libtool. The documentation has been turned into a man page, so
3926 there are now .1, .txt, and .html versions in /doc.
3928 5. Upgrades to pcregrep:
3929 (i) Added long-form option names like gnu grep.
3930 (ii) Added --help to list all options with an explanatory phrase.
3931 (iii) Added -r, --recursive to recurse into sub-directories.
3932 (iv) Added -f, --file to read patterns from a file.
3934 6. pcre_exec() was referring to its "code" argument before testing that
3935 argument for NULL (and giving an error if it was NULL).
3937 7. Upgraded Makefile.in to allow for compiling in a different directory from
3938 the source directory.
3940 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3941 options bits, the pointer it was passed was to an int instead of to an unsigned
3942 long int. This mattered only on 64-bit systems.
3944 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3945 generated) instead of pcre.in, which it its source. Also made the same change
3946 in several of the .c files.
3948 10. A new release of gcc defines printf() as a macro, which broke pcretest
3949 because it had an ifdef in the middle of a string argument for printf(). Fixed
3950 by using separate calls to printf().
3952 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3953 script, to force use of CR or LF instead of \n in the source. On non-Unix
3954 systems, the value can be set in config.h.
3956 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3957 absolute limit. Changed the text of the error message to make this clear, and
3958 likewise updated the man page.
3960 13. The limit of 99 on the number of capturing subpatterns has been removed.
3961 The new limit is 65535, which I hope will not be a "real" limit.
3964 Version 3.4 22-Aug-00
3965 ---------------------
3967 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3969 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3972 Version 3.3 01-Aug-00
3973 ---------------------
3975 1. If an octal character was given, but the value was greater than \377, it
3976 was not getting masked to the least significant bits, as documented. This could
3977 lead to crashes in some systems.
3979 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3980 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3982 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3983 These just pass their arguments on to (pcre_free)(), but they are provided
3984 because some uses of PCRE bind it to non-C systems that can call its functions,
3985 but cannot call free() or pcre_free() directly.
3987 4. Add "make test" as a synonym for "make check". Corrected some comments in
3988 the Makefile.
3990 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3991 Makefile.
3993 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3994 command called pgrep for grepping around the active processes.
3996 7. Added the beginnings of support for UTF-8 character strings.
3998 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3999 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4000 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4001 out for the ar command.)
4004 Version 3.2 12-May-00
4005 ---------------------
4007 This is purely a bug fixing release.
4009 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4010 of ZA. This was just one example of several cases that could provoke this bug,
4011 which was introduced by change 9 of version 2.00. The code for breaking
4012 infinite loops after an iteration that matches an empty string was't working
4013 correctly.
4015 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4016 when matched against abbab (for example). After matching an empty string, it
4017 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4018 caused it to match further down the string than it should.
4020 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4021 was there because it doesn't seem to be needed, and it causes trouble on some
4022 systems, as it is not a Standard C header. It has been removed.
4024 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4025 were reported on the Macintosh. The changes were from
4027 while ((c = *(++ptr)) != 0 && c != '\n');
4028 to
4029 while ((c = *(++ptr)) != 0 && c != '\n') ;
4031 Totally extraordinary, but if that's what it takes...
4033 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4034 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4035 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4036 assumes the way PCRE uses memmove() (always moving upwards).
4038 6. PCRE is being used in one environment where strchr() is not available. There
4039 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4040 faster code anyway.
4043 Version 3.1 09-Feb-00
4044 ---------------------
4046 The only change in this release is the fixing of some bugs in Makefile.in for
4047 the "install" target:
4049 (1) It was failing to install pcreposix.h.
4051 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4054 Version 3.0 01-Feb-00
4055 ---------------------
4057 1. Add support for the /+ modifier to perltest (to output $` like it does in
4058 pcretest).
4060 2. Add support for the /g modifier to perltest.
4062 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4063 matches null strings.
4065 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4066 pattern. Perl treats empty patterns specially - it reuses the most recent
4067 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4068 effect.
4070 5. The POSIX interface was broken in that it was just handing over the POSIX
4071 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4072 required a bigger vector, with some working space on the end. This means that
4073 the POSIX wrapper now has to get and free some memory, and copy the results.
4075 6. Added some simple autoconf support, placing the test data and the
4076 documentation in separate directories, re-organizing some of the
4077 information files, and making it build pcre-config (a GNU standard). Also added
4078 libtool support for building PCRE as a shared library, which is now the
4079 default.
4081 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4082 09 are not valid octal constants. Single digits will be used for minor values
4083 less than 10.
4085 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4086 existing programs that set these in the POSIX interface can use PCRE without
4087 modification.
4089 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4090 return all that pcre_info() returns, plus additional data. The pcre_info()
4091 function is retained for compatibility, but is considered to be obsolete.
4093 10. Added experimental recursion feature (?R) to handle one common case that
4094 Perl 5.6 will be able to do with (?p{...}).
4096 11. Added support for POSIX character classes like [:alpha:], which Perl is
4097 adopting.
4100 Version 2.08 31-Aug-99
4101 ----------------------
4103 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4104 trying to match at the startoffset position, but instead was moving forward to
4105 the next newline as if a previous match had failed.
4107 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4108 and could get into a loop if a null string was matched other than at the start
4109 of the subject.
4111 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4112 be distinguished at compile time, and for completeness also added PCRE_DATE.
4114 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4115 in GnuWin32 environments.
4118 Version 2.07 29-Jul-99
4119 ----------------------
4121 1. The documentation is now supplied in plain text form and HTML as well as in
4122 the form of man page sources.
4124 2. C++ compilers don't like assigning (void *) values to other pointer types.
4125 In particular this affects malloc(). Although there is no problem in Standard
4126 C, I've put in casts to keep C++ compilers happy.
4128 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4129 should be (const char *).
4131 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4132 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4133 However, I haven't made this a standard facility. The documentation doesn't
4134 mention it, and the Makefile doesn't support it.
4136 5. The Makefile now contains an "install" target, with editable destinations at
4137 the top of the file. The pcretest program is not installed.
4139 6. pgrep -V now gives the PCRE version number and date.
4141 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4142 causing the entire string to be ignored, instead of just the last character.
4144 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4145 non-matching string, it can take a very, very long time, even for strings of
4146 quite modest length, because of the nested recursion. PCRE now does better in
4147 some of these cases. It does this by remembering the last required literal
4148 character in the pattern, and pre-searching the subject to ensure it is present
4149 before running the real match. In other words, it applies a heuristic to detect
4150 some types of certain failure quickly, and in the above example, if presented
4151 with a string that has no trailing " it gives "no match" very quickly.
4153 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4154 other alternatives are tried instead.
4157 Version 2.06 09-Jun-99
4158 ----------------------
4160 1. Change pcretest's output for amount of store used to show just the code
4161 space, because the remainder (the data block) varies in size between 32-bit and
4162 64-bit systems.
4164 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4165 start matching at. This allows lookbehinds to work when searching for multiple
4166 occurrences in a string.
4168 3. Added additional options to pcretest for testing multiple occurrences:
4170 /+ outputs the rest of the string that follows a match
4171 /g loops for multiple occurrences, using the new startoffset argument
4172 /G loops for multiple occurrences by passing an incremented pointer
4174 4. PCRE wasn't doing the "first character" optimization for patterns starting
4175 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4176 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4177 the letter 'x'. On long subject strings, this gives a significant speed-up.
4180 Version 2.05 21-Apr-99
4181 ----------------------
4183 1. Changed the type of magic_number from int to long int so that it works
4184 properly on 16-bit systems.
4186 2. Fixed a bug which caused patterns starting with .* not to work correctly
4187 when the subject string contained newline characters. PCRE was assuming
4188 anchoring for such patterns in all cases, which is not correct because .* will
4189 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4190 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4191 must be retried after every newline in the subject.
4194 Version 2.04 18-Feb-99
4195 ----------------------
4197 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4198 computation of the store needed to hold the pattern was incorrect (too large).
4199 If such patterns were nested a few deep, this could multiply and become a real
4200 problem.
4202 2. Added /M option to pcretest to show the memory requirement of a specific
4203 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4205 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4206 compiled in such a way that the backtracking after subsequent failure was
4207 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4208 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4211 Version 2.03 02-Feb-99
4212 ----------------------
4214 1. Fixed typo and small mistake in man page.
4216 2. Added 4th condition (GPL supersedes if conflict) and created separate
4217 LICENCE file containing the conditions.
4219 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4220 Perl, that is the internal \ allows the delimiter to be included in the
4221 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4222 the final delimiter, add \ to the end of the pattern (to test the error).
4224 4. Added the convenience functions for extracting substrings after a successful
4225 match. Updated pcretest to make it able to test these functions.
4228 Version 2.02 14-Jan-99
4229 ----------------------
4231 1. Initialized the working variables associated with each extraction so that
4232 their saving and restoring doesn't refer to uninitialized store.
4234 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4235 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4236 fix the problem.
4238 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4239 calls, and wasn't printing the correct value for compiling calls. Increased the
4240 default value of LOOPREPEAT, and the number of significant figures in the
4241 times.
4243 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4245 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4246 a building problem on Windows NT with a FAT file system.
4249 Version 2.01 21-Oct-98
4250 ----------------------
4252 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4253 to character tables built by pcre_maketables() in the current locale. If NULL
4254 is passed, the default tables are used.
4257 Version 2.00 24-Sep-98
4258 ----------------------
4260 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4261 it any more.
4263 2. Allow quantification of (?>) groups, and make it work correctly.
4265 3. The first character computation wasn't working for (?>) groups.
4267 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4268 end of the subject) and add 5.005's \z, which really does match only at the
4269 very end of the subject.
4271 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4273 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4274 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4275 localized options. All options to pcre_study() were also removed.
4277 7. Add other new features from 5.005:
4279 $(?<= positive lookbehind
4280 $(?<! negative lookbehind
4281 (?imsx-imsx) added the unsetting capability
4282 such a setting is global if at outer level; local otherwise
4283 (?imsx-imsx:) non-capturing groups with option setting
4284 (?(cond)re|re) conditional pattern matching
4286 A backreference to itself in a repeated group matches the previous
4287 captured string.
4289 8. General tidying up of studying (both automatic and via "study")
4290 consequential on the addition of new assertions.
4292 9. As in 5.005, unlimited repeated groups that could match an empty substring
4293 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4294 runtime if any iteration does actually match an empty substring.
4296 10. Include the RunTest script in the distribution.
4298 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4299 discrepancies, some of which were old and were also with respect to 5.004. They
4300 have now been fixed.
4303 Version 1.09 28-Apr-98
4304 ----------------------
4306 1. A negated single character class followed by a quantifier with a minimum
4307 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4308 program crashes, or just wrong answers. This did not apply to negated classes
4309 containing more than one character, or to minima other than one.
4312 Version 1.08 27-Mar-98
4313 ----------------------
4315 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4317 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4318 latter must appear before anything that relies on it in the pattern.
4321 Version 1.07 16-Feb-98
4322 ----------------------
4324 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4325 repeat of a potentially empty string).
4328 Version 1.06 23-Jan-98
4329 ----------------------
4331 1. Added Markus Oberhumer's little patches for C++.
4333 2. Literal strings longer than 255 characters were broken.
4336 Version 1.05 23-Dec-97
4337 ----------------------
4339 1. Negated character classes containing more than one character were failing if
4340 PCRE_CASELESS was set at run time.
4343 Version 1.04 19-Dec-97
4344 ----------------------
4346 1. Corrected the man page, where some "const" qualifiers had been omitted.
4348 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4349 input syntax.
4351 3. Fixed memory leak which occurred when a regex with back references was
4352 matched with an offsets vector that wasn't big enough. The temporary memory
4353 that is used in this case wasn't being freed if the match failed.
4355 4. Tidied pcretest to ensure it frees memory that it gets.
4357 5. Temporary memory was being obtained in the case where the passed offsets
4358 vector was exactly big enough.
4360 6. Corrected definition of offsetof() from change 5 below.
4362 7. I had screwed up change 6 below and broken the rules for the use of
4363 setjmp(). Now fixed.
4366 Version 1.03 18-Dec-97
4367 ----------------------
4369 1. A erroneous regex with a missing opening parenthesis was correctly
4370 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4371 on some systems.
4373 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4374 it was reported that one broken compiler failed on the former because "code" is
4375 also an independent variable.
4377 3. The erroneous regex a[]b caused an array overrun reference.
4379 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4380 fail on data ending with that character. (It was going on too far, and checking
4381 the next character, typically a binary zero.) This was specific to the
4382 optimized code for single-character negative classes.
4384 5. Added a contributed patch from the TIN world which does the following:
4386 + Add an undef for memmove, in case the the system defines a macro for it.
4388 + Add a definition of offsetof(), in case there isn't one. (I don't know
4389 the reason behind this - offsetof() is part of the ANSI standard - but
4390 it does no harm).
4392 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4393 most of the places where whitespace preceded '#'. I have given up and
4394 allowed the remaining 2 cases to be at the margin.
4396 + Rename some variables in pcre to eliminate shadowing. This seems very
4397 pedantic, but does no harm, of course.
4399 6. Moved the call to setjmp() into its own function, to get rid of warnings
4400 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4402 7. Constructs such as \d{8,} were compiling into the equivalent of
4403 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4404 outcome, but in this particular case used more store than had been allocated,
4405 which caused the bug to be discovered because it threw up an internal error.
4407 8. The debugging code in both pcre and pcretest for outputting the compiled
4408 form of a regex was going wrong in the case of back references followed by
4409 curly-bracketed repeats.
4412 Version 1.02 12-Dec-97
4413 ----------------------
4415 1. Typos in pcre.3 and comments in the source fixed.
4417 2. Applied a contributed patch to get rid of places where it used to remove
4418 'const' from variables, and fixed some signed/unsigned and uninitialized
4419 variable warnings.
4421 3. Added the "runtest" target to Makefile.
4423 4. Set default compiler flag to -O2 rather than just -O.
4426 Version 1.01 19-Nov-97
4427 ----------------------
4429 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4430 like /([ab]*)*/, that is, for classes with more than one character in them.
4432 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4433 as /((?>a*))*/ (a PCRE_EXTRA facility).
4436 Version 1.00 18-Nov-97
4437 ----------------------
4439 1. Added compile-time macros to support systems such as SunOS4 which don't have
4440 memmove() or strerror() but have other things that can be used instead.
4442 2. Arranged that "make clean" removes the executables.
4445 Version 0.99 27-Oct-97
4446 ----------------------
4448 1. Fixed bug in code for optimizing classes with only one character. It was
4449 initializing a 32-byte map regardless, which could cause it to run off the end
4450 of the memory it had got.
4452 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4455 Version 0.98 22-Oct-97
4456 ----------------------
4458 1. Fixed bug in code for handling temporary memory usage when there are more
4459 back references than supplied space in the ovector. This could cause segfaults.
4462 Version 0.97 21-Oct-97
4463 ----------------------
4465 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4467 2. Optimized negated single characters not to use a bit map.
4469 3. Brought error texts together as macro definitions; clarified some of them;
4470 fixed one that was wrong - it said "range out of order" when it meant "invalid
4471 escape sequence".
4473 4. Changed some char * arguments to const char *.
4475 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4477 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4478 pcretest.
4481 Version 0.96 16-Oct-97
4482 ----------------------
4484 1. Added a simple "pgrep" utility to the distribution.
4486 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4487 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4488 where "ddd" means "one or more decimal digits".
4490 3. Fixed serious bug. If a pattern had a back reference, but the call to
4491 pcre_exec() didn't supply a large enough ovector to record the related
4492 identifying subpattern, the match always failed. PCRE now remembers the number
4493 of the largest back reference, and gets some temporary memory in which to save
4494 the offsets during matching if necessary, in order to ensure that
4495 backreferences always work.
4497 4. Increased the compatibility with Perl in a number of ways:
4499 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4500 to request this handling. The option can be set at compile or exec time.
4502 (b) $ matches before a terminating newline by default; an option
4503 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4504 mode). The option can be set at compile or exec time.
4506 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4507 the same as Perl's. If the decimal number it represents is less than 10
4508 or there aren't that many previous left capturing parentheses, an octal
4509 escape is read. Inside a character class, it's always an octal escape,
4510 even if it is a single digit.
4512 (d) An escaped but undefined alphabetic character is taken as a literal,
4513 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4514 escapes.
4516 (e) {0} is now permitted. (The previous item is removed from the compiled
4517 pattern).
4519 5. Changed all the names of code files so that the basic parts are no longer
4520 than 10 characters, and abolished the teeny "globals.c" file.
4522 6. Changed the handling of character classes; they are now done with a 32-byte
4523 bit map always.
4525 7. Added the -d and /D options to pcretest to make it possible to look at the
4526 internals of compilation without having to recompile pcre.
4529 Version 0.95 23-Sep-97
4530 ----------------------
4532 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4533 \x20 at the start of a run of normal characters. These were being treated as
4534 real characters, instead of the source characters being re-checked.
4537 Version 0.94 18-Sep-97
4538 ----------------------
4540 1. The functions are now thread-safe, with the caveat that the global variables
4541 containing pointers to malloc() and free() or alternative functions are the
4542 same for all threads.
4544 2. Get pcre_study() to generate a bitmap of initial characters for non-
4545 anchored patterns when this is possible, and use it if passed to pcre_exec().
4548 Version 0.93 15-Sep-97
4549 ----------------------
4551 1. /(b)|(:+)/ was computing an incorrect first character.
4553 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4554 but not actually doing anything yet.
4556 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4557 as Perl does (e.g. [-az] or [az-]).
4559 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4560 all possible positions.
4562 5. Split up into different modules to avoid including unneeded functions in a
4563 compiled binary. However, compile and exec are still in one module. The "study"
4564 function is split off.
4566 6. The character tables are now in a separate module whose source is generated
4567 by an auxiliary program - but can then be edited by hand if required. There are
4568 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4569 toupper() in the code.
4571 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4572 make them global. Abolish the function for setting them, as the caller can now
4573 set them directly.
4576 Version 0.92 11-Sep-97
4577 ----------------------
4579 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4580 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4582 2. Caseless matching was not working in character classes if the characters in
4583 the pattern were in upper case.
4585 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4587 4. Make PCRE_ANCHORED public and accept as a compile option.
4589 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4590 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4591 pass them.
4593 6. Give an error if bad option bits passed at compile or run time.
4595 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4596 pcretest to cause it to pass that flag.
4598 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4599 options, and the first character, if set.
4601 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4604 Version 0.91 10-Sep-97
4605 ----------------------
4607 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4608 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4610 2. PCRE was looping on encountering an indefinitely repeated back reference to
4611 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4612 Perl does - treats the match as successful.
4614 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5