ViewVC logotype

Contents of /code/branches/pcre16/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 832 - (show annotations)
Tue Dec 27 17:32:22 2011 UTC (7 years, 9 months ago) by ph10
File size: 224207 byte(s)
Fix back references with zero minimum repeat when the reference is unset.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.30
5 ------------
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
10 2. Fixed a bug in the code for calculating the fixed length of lookbehind
11 assertions.
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
27 Version 8.21 12-Dec-2011
28 ------------------------
30 1. Updating the JIT compiler.
32 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
33 are added as well.
35 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
36 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
37 calling _pcre_jit_exec. Some extra comments are added.
39 4. (*MARK) settings inside atomic groups that do not contain any capturing
40 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
41 was introduced by change 18 for 8.20.
43 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
44 ECMA-262 standard.
46 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
47 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
48 This bug was probably introduced by change 9 of 8.13.
50 7. While fixing 6 above, I noticed that a number of other items were being
51 incorrectly rejected as "not fixed length". This arose partly because newer
52 opcodes had not been added to the fixed-length checking code. I have (a)
53 corrected the bug and added tests for these items, and (b) arranged for an
54 error to occur if an unknown opcode is encountered while checking for fixed
55 length instead of just assuming "not fixed length". The items that were
56 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
57 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
58 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
60 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
61 being incorrectly compiled and would have given unpredicatble results.
63 9. A possessively repeated subpattern with minimum repeat count greater than
64 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
65 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
66 the first (A) could occur when it should not.
68 10. Add a cast and remove a redundant test from the code.
70 11. JIT should use pcre_malloc/pcre_free for allocation.
72 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
73 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
74 is anything other than /usr, -L is still shown).
76 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
78 14. Perl does not support \N without a following name in a [] class; PCRE now
79 also gives an error.
81 15. If a forward reference was repeated with an upper limit of around 2000,
82 it caused the error "internal error: overran compiling workspace". The
83 maximum number of forward references (including repeats) was limited by the
84 internal workspace, and dependent on the LINK_SIZE. The code has been
85 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
86 the default depends on LINK_SIZE. There is a new upper limit (for safety)
87 of around 200,000 forward references. While doing this, I also speeded up
88 the filling in of repeated forward references.
90 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
91 incorrectly expecting the subject to contain another "a" after the start.
93 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
94 in the match, the SKIP should be ignored. This was not happening; instead
95 the SKIP was being treated as NOMATCH. For patterns such as
96 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
97 tested.
99 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
100 now much more compatible with Perl, in particular in cases where the result
101 is a non-match for a non-anchored pattern. For example, if
102 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
103 "m", where previously it did not return a name. A side effect of this
104 change is that for partial matches, the last encountered mark name is
105 returned, as for non matches. A number of tests that were previously not
106 Perl-compatible have been moved into the Perl-compatible test files. The
107 refactoring has had the pleasing side effect of removing one argument from
108 the match() function, thus reducing its stack requirements.
110 19. If the /S+ option was used in pcretest to study a pattern using JIT,
111 subsequent uses of /S (without +) incorrectly behaved like /S+.
113 21. Retrieve executable code size support for the JIT compiler and fixing
114 some warnings.
116 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
117 not work when the shorter character appeared right at the end of the
118 subject string.
120 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
121 systems.
123 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
124 output it when the /M option is used in pcretest.
126 25. The CheckMan script was not being included in the distribution. Also, added
127 an explicit "perl" to run Perl scripts from the PrepareRelease script
128 because this is reportedly needed in Windows.
130 26. If study data was being save in a file and studying had not found a set of
131 "starts with" bytes for the pattern, the data written to the file (though
132 never used) was taken from uninitialized memory and so caused valgrind to
133 complain.
135 27. Updated RunTest.bat as provided by Sheri Pierce.
137 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
139 29. Computation of memory usage for the table of capturing group names was
140 giving an unnecessarily large value.
143 Version 8.20 21-Oct-2011
144 ------------------------
146 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
147 a POSIX class. After further experiments with Perl, which convinced me that
148 Perl has bugs and confusions, a closing square bracket is no longer allowed
149 in a POSIX name. This bug also affected patterns with classes that started
150 with full stops.
152 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
153 captured substring, but while checking the failing first alternative,
154 substring 1 is temporarily captured. If the output vector supplied to
155 pcre_exec() was not big enough for this capture, the yield of the function
156 was still zero ("insufficient space for captured substrings"). This cannot
157 be totally fixed without adding another stack variable, which seems a lot
158 of expense for a edge case. However, I have improved the situation in cases
159 such as /(a)(b)x|abc/ matched against "abc", where the return code
160 indicates that fewer than the maximum number of slots in the ovector have
161 been set.
163 3. Related to (2) above: when there are more back references in a pattern than
164 slots in the output vector, pcre_exec() uses temporary memory during
165 matching, and copies in the captures as far as possible afterwards. It was
166 using the entire output vector, but this conflicts with the specification
167 that only 2/3 is used for passing back captured substrings. Now it uses
168 only the first 2/3, for compatibility. This is, of course, another edge
169 case.
171 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
172 main code base, and can be used by building with --enable-jit. When this is
173 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
174 runtime --no-jit option is given.
176 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
177 ovector, the return from the function was zero, implying that there were
178 other matches that did not fit. The correct "exactly full" value is now
179 returned.
181 6. If a subpattern that was called recursively or as a subroutine contained
182 (*PRUNE) or any other control that caused it to give a non-standard return,
183 invalid errors such as "Error -26 (nested recursion at the same subject
184 position)" or even infinite loops could occur.
186 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
187 computing the minimum length on reaching *ACCEPT, and so ended up with the
188 wrong value of 1 rather than 0. Further investigation indicates that
189 computing a minimum subject length in the presence of *ACCEPT is difficult
190 (think back references, subroutine calls), and so I have changed the code
191 so that no minimum is registered for a pattern that contains *ACCEPT.
193 8. If (*THEN) was present in the first (true) branch of a conditional group,
194 it was not handled as intended. [But see 16 below.]
196 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
197 Sheri Pierce.
199 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
200 the first byte in a match must be "a".
202 11. Change 17 for 8.13 increased the recursion depth for patterns like
203 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
204 pattern contains any instances of (*THEN). If it does not, the old
205 optimizations are restored. It would be nice to do this on a per-group
206 basis, but at the moment that is not feasible.
208 12. In some environments, the output of pcretest -C is CRLF terminated. This
209 broke RunTest's code that checks for the link size. A single white space
210 character after the value is now allowed for.
212 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
213 For "fr", it uses the Windows-specific input and output files.
215 14. If (*THEN) appeared in a group that was called recursively or as a
216 subroutine, it did not work as intended. [But see next item.]
218 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
219 pattern fragments (but not containing any | characters). If A and B are
220 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
221 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
222 D. In other words, Perl considers parentheses that do not contain any |
223 characters to be part of a surrounding alternative, whereas PCRE was
224 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
225 differently. PCRE now behaves in the same way as Perl, except in the case
226 of subroutine/recursion calls such as (?1) which have in any case always
227 been different (but PCRE had them first :-).
229 16. Related to 15 above: Perl does not treat the | in a conditional group as
230 creating alternatives. Such a group is treated in the same way as an
231 ordinary group without any | characters when processing (*THEN). PCRE has
232 been changed to match Perl's behaviour.
234 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
235 RunGrepTest script failed.
237 18. Change 22 for version 13 caused atomic groups to use more stack. This is
238 inevitable for groups that contain captures, but it can lead to a lot of
239 stack use in large patterns. The old behaviour has been restored for atomic
240 groups that do not contain any capturing parentheses.
242 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
243 suppress the check for a minimum subject length at run time. (If it was
244 given to pcre_exec() or pcre_dfa_exec() it did work.)
246 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
247 fail to work when decoding hex characters in data strings in EBCDIC
248 environments.
250 21. It appears that in at least one Mac OS environment, the isxdigit() function
251 is implemented as a macro that evaluates to its argument more than once,
252 contravening the C 90 Standard (I haven't checked a later standard). There
253 was an instance in pcretest which caused it to go wrong when processing
254 \x{...} escapes in subject strings. The has been rewritten to avoid using
255 things like p++ in the argument of isxdigit().
258 Version 8.13 16-Aug-2011
259 ------------------------
261 1. The Unicode data tables have been updated to Unicode 6.0.0.
263 2. Two minor typos in pcre_internal.h have been fixed.
265 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
266 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
267 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
269 4. There were a number of related bugs in the code for matching backrefences
270 caselessly in UTF-8 mode when codes for the characters concerned were
271 different numbers of bytes. For example, U+023A and U+2C65 are an upper
272 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
273 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
274 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
275 2-byte code at the end of the subject (it thought there wasn't enough data
276 left).
278 5. Comprehensive information about what went wrong is now returned by
279 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
280 as the output vector has at least 2 elements. The offset of the start of
281 the failing character and a reason code are placed in the vector.
283 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
284 now returned is for the first byte of the failing character, instead of the
285 last byte inspected. This is an incompatible change, but I hope it is small
286 enough not to be a problem. It makes the returned offset consistent with
287 pcre_exec() and pcre_dfa_exec().
289 7. pcretest now gives a text phrase as well as the error number when
290 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
291 failure, the offset and reason code are output.
293 8. When \R was used with a maximizing quantifier it failed to skip backwards
294 over a \r\n pair if the subsequent match failed. Instead, it just skipped
295 back over a single character (\n). This seems wrong (because it treated the
296 two characters as a single entity when going forwards), conflicts with the
297 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
298 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
299 has been changed.
301 9. Some internal refactoring has changed the processing so that the handling
302 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
303 time (the PCRE_DOTALL option was changed this way some time ago: version
304 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
305 which was always a bit of a fudge. It also means that there is one less
306 argument for the match() function, which reduces its stack requirements
307 slightly. This change also fixes an incompatibility with Perl: the pattern
308 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
310 10. More internal refactoring has drastically reduced the number of recursive
311 calls to match() for possessively repeated groups such as (abc)++ when
312 using pcre_exec().
314 11. While implementing 10, a number of bugs in the handling of groups were
315 discovered and fixed:
317 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
318 (a|)*(?1) gave a compile-time internal error.
319 ((a|)+)+ did not notice that the outer group could match an empty string.
320 (^a|^)+ was not marked as anchored.
321 (.*a|.*)+ was not marked as matching at start or after a newline.
323 12. Yet more internal refactoring has removed another argument from the match()
324 function. Special calls to this function are now indicated by setting a
325 value in a variable in the "match data" data block.
327 13. Be more explicit in pcre_study() instead of relying on "default" for
328 opcodes that mean there is no starting character; this means that when new
329 ones are added and accidentally left out of pcre_study(), testing should
330 pick them up.
332 14. The -s option of pcretest has been documented for ages as being an old
333 synonym of -m (show memory usage). I have changed it to mean "force study
334 for every regex", that is, assume /S for every regex. This is similar to -i
335 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
336 using it. It makes it easier to run collections of tests with and without
337 study enabled, and thereby test pcre_study() more easily. All the standard
338 tests are now run with and without -s (but some patterns can be marked as
339 "never study" - see 20 below).
341 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
342 restoration of the capturing data to the outer values was not happening
343 correctly.
345 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
346 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
347 pattern had matched an empty string, and so incorrectly returned a no
348 match.
350 17. There was optimizing code for the last branch of non-capturing parentheses,
351 and also for the obeyed branch of a conditional subexpression, which used
352 tail recursion to cut down on stack usage. Unfortunately, now that there is
353 the possibility of (*THEN) occurring in these branches, tail recursion is
354 no longer possible because the return has to be checked for (*THEN). These
355 two optimizations have therefore been removed. [But see 8.20/11 above.]
357 18. If a pattern containing \R was studied, it was assumed that \R always
358 matched two bytes, thus causing the minimum subject length to be
359 incorrectly computed because \R can also match just one byte.
361 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
362 was incorrectly computed.
364 20. If /S is present twice on a test pattern in pcretest input, it now
365 *disables* studying, thereby overriding the use of -s on the command line
366 (see 14 above). This is necessary for one or two tests to keep the output
367 identical in both cases.
369 21. When (*ACCEPT) was used in an assertion that matched an empty string and
370 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
372 22. When an atomic group that contained a capturing parenthesis was
373 successfully matched, but the branch in which it appeared failed, the
374 capturing was not being forgotten if a higher numbered group was later
375 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
376 group 1 to "a", when in fact it should be unset. This applied to multi-
377 branched capturing and non-capturing groups, repeated or not, and also to
378 positive assertions (capturing in negative assertions does not happen
379 in PCRE) and also to nested atomic groups.
381 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
382 subject after a captured substring, to make it easier to tell which of a
383 number of identical substrings has been captured.
385 24. The way atomic groups are processed by pcre_exec() has been changed so that
386 if they are repeated, backtracking one repetition now resets captured
387 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
388 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
389 "aaa". Previously, it would have been "a". As part of this code
390 refactoring, the way recursive calls are handled has also been changed.
392 25. If an assertion condition captured any substrings, they were not passed
393 back unless some other capturing happened later. For example, if
394 (?(?=(a))a) was matched against "a", no capturing was returned.
396 26. When studying a pattern that contained subroutine calls or assertions,
397 the code for finding the minimum length of a possible match was handling
398 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
399 group 1 called group 2 while simultaneously a separate group 2 called group
400 1). A stack overflow occurred in this case. I have fixed this by limiting
401 the recursion depth to 10.
403 27. Updated RunTest.bat in the distribution to the version supplied by Tom
404 Fortmann. This supports explicit test numbers on the command line, and has
405 argument validation and error reporting.
407 28. An instance of \X with an unlimited repeat could fail if at any point the
408 first character it looked at was a mark character.
410 29. Some minor code refactoring concerning Unicode properties and scripts
411 should reduce the stack requirement of match() slightly.
413 30. Added the '=' option to pcretest to check the setting of unused capturing
414 slots at the end of the pattern, which are documented as being -1, but are
415 not included in the return count.
417 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
418 compiled something random. Now it gives a compile-time error (as does
419 Perl).
421 32. A *MARK encountered during the processing of a positive assertion is now
422 recorded and passed back (compatible with Perl).
424 33. If --only-matching or --colour was set on a pcregrep call whose pattern
425 had alternative anchored branches, the search for a second match in a line
426 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
427 matched the line "0102" twice. The same bug affected patterns that started
428 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
429 twice.
431 34. Previously, PCRE did not allow quantification of assertions. However, Perl
432 does, and because of capturing effects, quantifying parenthesized
433 assertions may at times be useful. Quantifiers are now allowed for
434 parenthesized assertions.
436 35. A minor code tidy in pcre_compile() when checking options for \R usage.
438 36. \g was being checked for fancy things in a character class, when it should
439 just be a literal "g".
441 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
442 appearance of a nested POSIX class supersedes an apparent external class.
443 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
444 unescaped square brackets may also appear as part of class names. For
445 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
446 more like Perl. (But see 8.20/1 above.)
448 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
449 was because it thought it was \N{name}, which is not supported).
451 39. Add minix to OS list not supporting the -S option in pcretest.
453 40. PCRE tries to detect cases of infinite recursion at compile time, but it
454 cannot analyze patterns in sufficient detail to catch mutual recursions
455 such as ((?1))((?2)). There is now a runtime test that gives an error if a
456 subgroup is called recursively as a subpattern for a second time at the
457 same position in the subject string. In previous releases this might have
458 been caught by the recursion limit, or it might have run out of stack.
460 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
461 happen only once. PCRE was, however incorrectly giving a compile time error
462 "recursive call could loop indefinitely" because it cannot analyze the
463 pattern in sufficient detail. The compile time test no longer happens when
464 PCRE is compiling a conditional subpattern, but actual runaway loops are
465 now caught at runtime (see 40 above).
467 42. It seems that Perl allows any characters other than a closing parenthesis
468 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
469 has been changed to be the same.
471 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
472 as not to get warnings when autogen.sh is called. Also changed
473 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
475 44. To help people who use pcregrep to scan files containing exceedingly long
476 lines, the following changes have been made:
478 (a) The default value of the buffer size parameter has been increased from
479 8K to 20K. (The actual buffer used is three times this size.)
481 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
482 PCRE is built.
484 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
485 to be set at run time.
487 (d) Numerical values in pcregrep options can be followed by K or M, for
488 example --buffer-size=50K.
490 (e) If a line being scanned overflows pcregrep's buffer, an error is now
491 given and the return code is set to 2.
493 45. Add a pointer to the latest mark to the callout data block.
495 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
496 partial match of an empty string instead of no match. This was specific to
497 the use of ".".
499 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
500 complete match instead of a partial match. This bug was dependent on both
501 the PCRE_UTF8 and PCRE_DOTALL options being set.
503 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
504 starting byte set, because \b was not being ignored.
507 Version 8.12 15-Jan-2011
508 ------------------------
510 1. Fixed some typos in the markup of the man pages, and wrote a script that
511 checks for such things as part of the documentation building process.
513 2. On a big-endian 64-bit system, pcregrep did not correctly process the
514 --match-limit and --recursion-limit options (added for 8.11). In
515 particular, this made one of the standard tests fail. (The integer value
516 went into the wrong half of a long int.)
518 3. If the --colour option was given to pcregrep with -v (invert match), it
519 did strange things, either producing crazy output, or crashing. It should,
520 of course, ignore a request for colour when reporting lines that do not
521 match.
523 4. Another pcregrep bug caused similar problems if --colour was specified with
524 -M (multiline) and the pattern match finished with a line ending.
526 5. In pcregrep, when a pattern that ended with a literal newline sequence was
527 matched in multiline mode, the following line was shown as part of the
528 match. This seems wrong, so I have changed it.
530 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
531 the check for further matches in the same line (so they could be coloured)
532 to overrun the end of the current line. If another match was found, it was
533 incorrectly shown (and then shown again when found in the next line).
535 7. If pcregrep was compiled under Windows, there was a reference to the
536 function pcregrep_exit() before it was defined. I am assuming this was
537 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
538 reported by a user. I've moved the definition above the reference.
541 Version 8.11 10-Dec-2010
542 ------------------------
544 1. (*THEN) was not working properly if there were untried alternatives prior
545 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
546 backtracked to try for "b" instead of moving to the next alternative branch
547 at the same level (in this case, to look for "c"). The Perl documentation
548 is clear that when (*THEN) is backtracked onto, it goes to the "next
549 alternative in the innermost enclosing group".
551 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
552 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
553 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
554 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
555 (*THEN).
557 3. If \s appeared in a character class, it removed the VT character from
558 the class, even if it had been included by some previous item, for example
559 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
560 of \s, but is part of the POSIX "space" class.)
562 4. A partial match never returns an empty string (because you can always
563 match an empty string at the end of the subject); however the checking for
564 an empty string was starting at the "start of match" point. This has been
565 changed to the "earliest inspected character" point, because the returned
566 data for a partial match starts at this character. This means that, for
567 example, /(?<=abc)def/ gives a partial match for the subject "abc"
568 (previously it gave "no match").
570 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
571 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
572 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
573 has an implication that the given string is incomplete (because a partial
574 match is preferred over a full match). For this reason, these items now
575 give a partial match in this situation. [Aside: previously, the one case
576 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
577 match rather than a full match, which was wrong by the old rules, but is
578 now correct.]
580 6. There was a bug in the handling of #-introduced comments, recognized when
581 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
582 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
583 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
584 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
585 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
586 places in pcre_compile().
588 7. Related to (6) above, when pcre_compile() was skipping #-introduced
589 comments when looking ahead for named forward references to subpatterns,
590 the only newline sequence it recognized was NL. It now handles newlines
591 according to the set newline convention.
593 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
594 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
595 cater for a lack of strerror(). These oversights have been fixed.
597 9. Added --match-limit and --recursion-limit to pcregrep.
599 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
601 11. When the -o option was used, pcregrep was setting a return code of 1, even
602 when matches were found, and --line-buffered was not being honoured.
604 12. Added an optional parentheses number to the -o and --only-matching options
605 of pcregrep.
607 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
608 can match an empty string. The code to do it in pcretest and pcredemo
609 needed fixing:
611 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
612 only one byte after an empty string match just before CRLF (this case
613 just got forgotten; "any" and "anycrlf" were OK).
615 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
616 mode when an empty string match preceded an ASCII character followed by
617 a non-ASCII character. (The code for advancing by one character rather
618 than one byte was nonsense.)
620 (c) The pcredemo.c sample program did not have any code at all to handle
621 the cases when CRLF is a valid newline sequence.
623 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
624 as a starting offset was within the subject string. There is now a new
625 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
626 negative or greater than the length of the string. In order to test this,
627 pcretest is extended to allow the setting of negative starting offsets.
629 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
630 starting offset points to the beginning of a UTF-8 character was
631 unnecessarily clumsy. I tidied it up.
633 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
634 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
636 17. Nobody had reported that the --include_dir option, which was added in
637 release 7.7 should have been called --include-dir (hyphen, not underscore)
638 for compatibility with GNU grep. I have changed it to --include-dir, but
639 left --include_dir as an undocumented synonym, and the same for
640 --exclude-dir, though that is not available in GNU grep, at least as of
641 release 2.5.4.
643 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
644 characters from a string of bytes) have been redefined so as not to use
645 loops, in order to improve performance in some environments. At the same
646 time, I abstracted some of the common code into auxiliary macros to save
647 repetition (this should not affect the compiled code).
649 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
650 compile-time error is now given if \c is not followed by an ASCII
651 character, that is, a byte less than 128. (In EBCDIC mode, the code is
652 different, and any byte value is allowed.)
654 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
655 START_OPTIMIZE option, which is now allowed at compile time - but just
656 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
657 to pcregrep and other applications that have no direct access to PCRE
658 options. The new /Y option in pcretest sets this option when calling
659 pcre_compile().
661 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
662 back references. Groups containing recursive back references were forced to
663 be atomic by that change, but in the case of named groups, the amount of
664 memory required was incorrectly computed, leading to "Failed: internal
665 error: code overflow". This has been fixed.
667 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
668 pcretest.c, to avoid build problems in some Borland environments.
671 Version 8.10 25-Jun-2010
672 ------------------------
674 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
675 THEN.
677 2. (*ACCEPT) was not working when inside an atomic group.
679 3. Inside a character class, \B is treated as a literal by default, but
680 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
681 causes the error). The code is unchanged, but I tidied the documentation.
683 4. Inside a character class, PCRE always treated \R and \X as literals,
684 whereas Perl faults them if its -w option is set. I have changed PCRE so
685 that it faults them when PCRE_EXTRA is set.
687 5. Added support for \N, which always matches any character other than
688 newline. (It is the same as "." when PCRE_DOTALL is not set.)
690 6. When compiling pcregrep with newer versions of gcc which may have
691 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
692 declared with attribute warn_unused_result" were given. Just casting the
693 result to (void) does not stop the warnings; a more elaborate fudge is
694 needed. I've used a macro to implement this.
696 7. Minor change to pcretest.c to avoid a compiler warning.
698 8. Added four artifical Unicode properties to help with an option to make
699 \s etc use properties (see next item). The new properties are: Xan
700 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
702 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
703 use Unicode properties. (*UCP) at the start of a pattern can be used to set
704 this option. Modified pcretest to add /W to test this facility. Added
705 REG_UCP to make it available via the POSIX interface.
707 10. Added --line-buffered to pcregrep.
709 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
710 studied, and the match started with a letter with a code point greater than
711 127 whose first byte was different to the first byte of the other case of
712 the letter, the other case of this starting letter was not recognized
713 (#976).
715 12. If a pattern that was studied started with a repeated Unicode property
716 test, for example, \p{Nd}+, there was the theoretical possibility of
717 setting up an incorrect bitmap of starting bytes, but fortunately it could
718 not have actually happened in practice until change 8 above was made (it
719 added property types that matched character-matching opcodes).
721 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
722 possible starting bytes for non-anchored patterns.
724 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
725 \R, and also a number of cases that involve Unicode properties, both
726 explicit and implicit when PCRE_UCP is set.
728 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
729 input, it could crash or give wrong results if characters with values
730 greater than 0xc0 were present in the subject string. (Detail: it assumed
731 UTF-8 input when processing these items.)
733 16. Added a lot of (int) casts to avoid compiler warnings in systems where
734 size_t is 64-bit (#991).
736 17. Added a check for running out of memory when PCRE is compiled with
737 --disable-stack-for-recursion (#990).
739 18. If the last data line in a file for pcretest does not have a newline on
740 the end, a newline was missing in the output.
742 19. The default pcre_chartables.c file recognizes only ASCII characters (values
743 less than 128) in its various bitmaps. However, there is a facility for
744 generating tables according to the current locale when PCRE is compiled. It
745 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
746 characters, are recognized by isspace() and therefore were getting set in
747 these tables, and indeed these tables seem to approximate to ISO 8859. This
748 caused a problem in UTF-8 mode when pcre_study() was used to create a list
749 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
750 which of course cannot start UTF-8 characters. I have changed the code so
751 that only real ASCII characters (less than 128) and the correct starting
752 bytes for UTF-8 encodings are set for characters greater than 127 when in
753 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
754 altogether.)
756 20. Added the /T option to pcretest so as to be able to run tests with non-
757 standard character tables, thus making it possible to include the tests
758 used for 19 above in the standard set of tests.
760 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
761 reference to a subpattern the other side of a comment that contains an
762 opening parenthesis caused either an internal compiling error, or a
763 reference to the wrong subpattern.
766 Version 8.02 19-Mar-2010
767 ------------------------
769 1. The Unicode data tables have been updated to Unicode 5.2.0.
771 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
772 configured.
774 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
775 original author of that file, following a query about its status.
777 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
778 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
780 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
781 quantifier applied to a forward-referencing subroutine call, could compile
782 incorrect code or give the error "internal error: previously-checked
783 referenced subpattern not found".
785 6. Both MS Visual Studio and Symbian OS have problems with initializing
786 variables to point to external functions. For these systems, therefore,
787 pcre_malloc etc. are now initialized to local functions that call the
788 relevant global functions.
790 7. There were two entries missing in the vectors called coptable and poptable
791 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
792 I've fixed the data, and added a kludgy way of testing at compile time that
793 the lengths are correct (equal to the number of opcodes).
795 8. Following on from 7, I added a similar kludge to check the length of the
796 eint vector in pcreposix.c.
798 9. Error texts for pcre_compile() are held as one long string to avoid too
799 much relocation at load time. To find a text, the string is searched,
800 counting zeros. There was no check for running off the end of the string,
801 which could happen if a new error number was added without updating the
802 string.
804 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
806 11. \K was not working if it appeared in an atomic group or in a group that
807 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
808 \K is "not well defined" if used in an assertion. PCRE now accepts it if
809 the assertion is positive, but not if it is negative.
811 12. Change 11 fortuitously reduced the size of the stack frame used in the
812 "match()" function of pcre_exec.c by one pointer. Forthcoming
813 implementation of support for (*MARK) will need an extra pointer on the
814 stack; I have reserved it now, so that the stack frame size does not
815 decrease.
817 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
818 item in branch that calls a recursion is a subroutine call - as in the
819 second branch in the above example - was incorrectly given the compile-
820 time error "recursive call could loop indefinitely" because pcre_compile()
821 was not correctly checking the subroutine for matching a non-empty string.
823 14. The checks for overrunning compiling workspace could trigger after an
824 overrun had occurred. This is a "should never occur" error, but it can be
825 triggered by pathological patterns such as hundreds of nested parentheses.
826 The checks now trigger 100 bytes before the end of the workspace.
828 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
831 Version 8.01 19-Jan-2010
832 ------------------------
834 1. If a pattern contained a conditional subpattern with only one branch (in
835 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
836 computed the wrong minimum data length (which is of course zero for such
837 subpatterns). This could cause incorrect "no match" results.
839 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
840 the pattern is reset in the first branch, pcre_compile() failed with
841 "internal error: code overflow at offset...". This happened only when
842 the reset was to the original external option setting. (An optimization
843 abstracts leading options settings into an external setting, which was the
844 cause of this.)
846 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
847 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
848 assertion pattern did not match (meaning that the assertion was true), it
849 was incorrectly treated as false if the SKIP had been reached during the
850 matching. This also applied to assertions used as conditions.
852 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
853 assertion subpattern, including such a pattern used as a condition,
854 unpredictable results occurred, instead of the error return
857 5. The C++ GlobalReplace function was not working like Perl for the special
858 situation when an empty string is matched. It now does the fancy magic
859 stuff that is necessary.
861 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
862 removed. (These were left over from very, very early versions of PCRE.)
864 7. Some cosmetic changes to the code to make life easier when compiling it
865 as part of something else:
867 (a) Change DEBUG to PCRE_DEBUG.
869 (b) In pcre_compile(), rename the member of the "branch_chain" structure
870 called "current" as "current_branch", to prevent a collision with the
871 Linux macro when compiled as a kernel module.
873 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
874 prevent a collision with the Linux macro when compiled as a kernel
875 module.
877 8. In pcre_compile() there are some checks for integer overflows that used to
878 cast potentially large values to (double). This has been changed to that
879 when building, a check for int64_t is made, and if it is found, it is used
880 instead, thus avoiding the use of floating point arithmetic. (There is no
881 other use of FP in PCRE.) If int64_t is not found, the fallback is to
882 double.
884 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
885 2005 (difference between two addresses compared to an unsigned value).
887 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
888 custom one, because of the following reported problem in Windows:
890 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
891 under Win32.
892 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
893 therefore missing the function definition.
894 - The compiler thus generates a "C" signature for the test function.
895 - The linker fails to find the "C" function.
896 - PCRE fails to configure if asked to do so against libbz2.
898 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
899 messages were output:
901 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
902 rerunning libtoolize, to keep the correct libtool macros in-tree.
903 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
905 I have done both of these things.
907 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
908 most of the time, it *can* run out if it is given a pattern that contains a
909 runaway infinite recursion. I updated the discussion in the pcrestack man
910 page.
912 13. Now that we have gone to the x.xx style of version numbers, the minor
913 version may start with zero. Using 08 or 09 is a bad idea because users
914 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
915 interpreted as invalid octal numbers. I've updated the previous comment in
916 configure.ac, and also added a check that gives an error if 08 or 09 are
917 used.
919 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
920 causing partial matching to fail when the end of the subject matched \W
921 in a UTF-8 pattern where \W was quantified with a minimum of 3.
923 15. There were some discrepancies between the declarations in pcre_internal.h
924 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
925 their definitions. The declarations used "const uschar *" and the
926 definitions used USPTR. Even though USPTR is normally defined as "const
927 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
928 reported that: "This difference in casting confuses some C++ compilers, for
929 example, SunCC recognizes above declarations as different functions and
930 generates broken code for hbpcre." I have changed the declarations to use
931 USPTR.
933 16. GNU libtool is named differently on some systems. The autogen.sh script now
934 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
935 (FreeBSD).
937 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
938 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
939 comment: "Figure out how to create a longlong from a string: strtoll and
940 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
941 instance, but it only takes 2 args instead of 3!"
943 18. A subtle bug concerned with back references has been fixed by a change of
944 specification, with a corresponding code fix. A pattern such as
945 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
946 refers, was giving matches when it shouldn't. For example, xa=xaaa would
947 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
948 same bug. Such groups have to be quantified to be useful, or contained
949 inside another quantified group. (If there's no repetition, the reference
950 can never match.) The problem arises because, having left the group and
951 moved on to the rest of the pattern, a later failure that backtracks into
952 the group uses the captured value from the final iteration of the group
953 rather than the correct earlier one. I have fixed this in PCRE by forcing
954 any group that contains a reference to itself to be an atomic group; that
955 is, there cannot be any backtracking into it once it has completed. This is
956 similar to recursive and subroutine calls.
959 Version 8.00 19-Oct-09
960 ----------------------
962 1. The table for translating pcre_compile() error codes into POSIX error codes
963 was out-of-date, and there was no check on the pcre_compile() error code
964 being within the table. This could lead to an OK return being given in
965 error.
967 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
968 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
969 in a Windows environment.
971 3. The pcregrep --count option prints the count for each file even when it is
972 zero, as does GNU grep. However, pcregrep was also printing all files when
973 --files-with-matches was added. Now, when both options are given, it prints
974 counts only for those files that have at least one match. (GNU grep just
975 prints the file name in this circumstance, but including the count seems
976 more useful - otherwise, why use --count?) Also ensured that the
977 combination -clh just lists non-zero counts, with no names.
979 4. The long form of the pcregrep -F option was incorrectly implemented as
980 --fixed_strings instead of --fixed-strings. This is an incompatible change,
981 but it seems right to fix it, and I didn't think it was worth preserving
982 the old behaviour.
984 5. The command line items --regex=pattern and --regexp=pattern were not
985 recognized by pcregrep, which required --regex pattern or --regexp pattern
986 (with a space rather than an '='). The man page documented the '=' forms,
987 which are compatible with GNU grep; these now work.
989 6. No libpcreposix.pc file was created for pkg-config; there was just
990 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
992 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
993 when UCP support is not needed, by modifying the Python script that
994 generates it from Unicode data files. This should not matter if the module
995 is correctly used as a library, but I received one complaint about 50K of
996 unwanted data. My guess is that the person linked everything into his
997 program rather than using a library. Anyway, it does no harm.
999 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1000 was a minimum greater than 1 for a wide character in a possessive
1001 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1002 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1003 character. Chaos in the form of incorrect output or a compiling loop could
1004 result.
1006 9. The restrictions on what a pattern can contain when partial matching is
1007 requested for pcre_exec() have been removed. All patterns can now be
1008 partially matched by this function. In addition, if there are at least two
1009 slots in the offset vector, the offset of the earliest inspected character
1010 for the match and the offset of the end of the subject are set in them when
1011 PCRE_ERROR_PARTIAL is returned.
1013 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1014 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1015 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1016 and may be more useful for multi-segment matching.
1018 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1019 used to be given if ever the end of the subject was reached; now it is
1020 given only if matching could not proceed because another character was
1021 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1022 string "Z", which now yields "no match" instead of "partial match". In the
1023 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1024 final character ended with (*FAIL).
1026 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1027 if the pattern had a "must contain" character that was already found in the
1028 earlier partial match, unless partial matching was again requested. For
1029 example, with the pattern /dog.(body)?/, the "must contain" character is
1030 "g". If the first part-match was for the string "dog", restarting with
1031 "sbody" failed. This bug has been fixed.
1033 13. The string returned by pcre_dfa_exec() after a partial match has been
1034 changed so that it starts at the first inspected character rather than the
1035 first character of the match. This makes a difference only if the pattern
1036 starts with a lookbehind assertion or \b or \B (\K is not supported by
1037 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1038 matching functions compatible, and I think it's the right thing to do.
1040 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1041 so that the demonstration program is easily available in environments where
1042 PCRE has not been installed from source.
1044 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1045 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1046 library.
1048 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1049 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1050 is not the first non-POSIX option to be added. Clearly some people find
1051 these options useful.
1053 17. If a caller to the POSIX matching function regexec() passes a non-zero
1054 value for nmatch with a NULL value for pmatch, the value of
1055 nmatch is forced to zero.
1057 18. RunGrepTest did not have a test for the availability of the -u option of
1058 the diff command, as RunTest does. It now checks in the same way as
1059 RunTest, and also checks for the -b option.
1061 19. If an odd number of negated classes containing just a single character
1062 interposed, within parentheses, between a forward reference to a named
1063 subpattern and the definition of the subpattern, compilation crashed with
1064 an internal error, complaining that it could not find the referenced
1065 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1066 [The bug was that it was starting one character too far in when skipping
1067 over the character class, thus treating the ] as data rather than
1068 terminating the class. This meant it could skip too much.]
1070 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1071 /g option in pcretest when the pattern contains \K, which makes it possible
1072 to have an empty string match not at the start, even when the pattern is
1073 anchored. Updated pcretest and pcredemo to use this option.
1075 21. If the maximum number of capturing subpatterns in a recursion was greater
1076 than the maximum at the outer level, the higher number was returned, but
1077 with unset values at the outer level. The correct (outer level) value is
1078 now given.
1080 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1081 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1082 make it do so. The string so far is captured, making this feature
1083 compatible with Perl.
1085 23. The tests have been re-organized, adding tests 11 and 12, to make it
1086 possible to check the Perl 5.10 features against Perl 5.10.
1088 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1089 pattern matches a fixed length string. PCRE did not allow this; now it
1090 does. Neither allows recursion.
1092 25. I finally figured out how to implement a request to provide the minimum
1093 length of subject string that was needed in order to match a given pattern.
1094 (It was back references and recursion that I had previously got hung up
1095 on.) This code has now been added to pcre_study(); it finds a lower bound
1096 to the length of subject needed. It is not necessarily the greatest lower
1097 bound, but using it to avoid searching strings that are too short does give
1098 some useful speed-ups. The value is available to calling programs via
1099 pcre_fullinfo().
1101 26. While implementing 25, I discovered to my embarrassment that pcretest had
1102 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1103 study optimizations had never been tested with that matching function.
1104 Oops. What is worse, even when it was passed study data, there was a bug in
1105 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1106 were also very few tests of studied patterns with pcre_dfa_exec().
1108 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1109 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1110 on the other side of the coin, they are no longer allowed to have different
1111 names, because these cannot be distinguished in PCRE, and this has caused
1112 confusion. (This is a difference from Perl.)
1114 28. When duplicate subpattern names are present (necessarily with different
1115 numbers, as required by 27 above), and a test is made by name in a
1116 conditional pattern, either for a subpattern having been matched, or for
1117 recursion in such a pattern, all the associated numbered subpatterns are
1118 tested, and the overall condition is true if the condition is true for any
1119 one of them. This is the way Perl works, and is also more like the way
1120 testing by number works.
1123 Version 7.9 11-Apr-09
1124 ---------------------
1126 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1127 (pcretest), all targets were linked against these libraries. This included
1128 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1129 libraries. This caused unwanted dependencies to be created. This problem
1130 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1131 pcretest is linked with readline.
1133 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1134 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1135 moved outside it again, because FALSE and TRUE are already defined in AIX,
1136 but BOOL is not.
1138 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1139 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1141 4. The pcregrep documentation said spaces were inserted as well as colons (or
1142 hyphens) following file names and line numbers when outputting matching
1143 lines. This is not true; no spaces are inserted. I have also clarified the
1144 wording for the --colour (or --color) option.
1146 5. In pcregrep, when --colour was used with -o, the list of matching strings
1147 was not coloured; this is different to GNU grep, so I have changed it to be
1148 the same.
1150 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1151 each matching line was coloured. Now it goes on to look for further matches
1152 of any of the test patterns, which is the same behaviour as GNU grep.
1154 7. A pattern that could match an empty string could cause pcregrep to loop; it
1155 doesn't make sense to accept an empty string match in pcregrep, so I have
1156 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1157 seems to be how GNU grep behaves.
1159 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1160 start or after a newline", because the conditional assertion was not being
1161 correctly handled. The rule now is that both the assertion and what follows
1162 in the first alternative must satisfy the test.
1164 9. If auto-callout was enabled in a pattern with a conditional group whose
1165 condition was an assertion, PCRE could crash during matching, both with
1166 pcre_exec() and pcre_dfa_exec().
1168 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1169 used for matching.
1171 11. Unicode property support in character classes was not working for
1172 characters (bytes) greater than 127 when not in UTF-8 mode.
1174 12. Added the -M command line option to pcretest.
1176 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1178 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1180 16. Added comments and documentation about mis-use of no_arg in the C++
1181 wrapper.
1183 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1184 from Martin Jerabek that uses macro names for all relevant character and
1185 string constants.
1187 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1188 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1189 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1190 these, but not everybody uses configure.
1192 19. A conditional group that had only one branch was not being correctly
1193 recognized as an item that could match an empty string. This meant that an
1194 enclosing group might also not be so recognized, causing infinite looping
1195 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1196 with the subject "ab", where knowledge that the repeated group can match
1197 nothing is needed in order to break the loop.
1199 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1200 exec(), but without supplying a callout function, matching went wrong.
1202 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1203 leak if the size of the offset vector was greater than 30. When the vector
1204 is smaller, the saved offsets during recursion go onto a local stack
1205 vector, but for larger vectors malloc() is used. It was failing to free
1206 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1207 error, in fact).
1209 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1210 heapframe that is used only when UTF-8 support is enabled. This caused no
1211 problem, but was untidy.
1213 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1214 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1215 included within another project.
1217 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1218 slightly modified by me:
1220 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1221 not building pcregrep.
1223 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1224 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1226 25. Forward references, both numeric and by name, in patterns that made use of
1227 duplicate group numbers, could behave incorrectly or give incorrect errors,
1228 because when scanning forward to find the reference group, PCRE was not
1229 taking into account the duplicate group numbers. A pattern such as
1230 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1232 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1233 the feature of a custom pointer more persuasive (as requested by a user).
1235 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1236 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1237 pre-defined.
1239 28. Added support for (*UTF8) at the start of a pattern.
1241 29. Arrange for flags added by the "release type" setting in CMake to be shown
1242 in the configuration summary.
1245 Version 7.8 05-Sep-08
1246 ---------------------
1248 1. Replaced UCP searching code with optimized version as implemented for Ad
1249 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1250 stage table and inline lookup instead of a function, giving speed ups of 2
1251 to 5 times on some simple patterns that I tested. Permission was given to
1252 distribute the MultiStage2.py script that generates the tables (it's not in
1253 the tarball, but is in the Subversion repository).
1255 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1256 scripts.
1258 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1259 a group with a zero qualifier. The result of the study could be incorrect,
1260 or the function might crash, depending on the pattern.
1262 4. Caseless matching was not working for non-ASCII characters in back
1263 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1264 It now works when Unicode Property Support is available.
1266 5. In pcretest, an escape such as \x{de} in the data was always generating
1267 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1268 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1269 truncation.
1271 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1273 7. Added two (int) casts to pcregrep when printing the difference of two
1274 pointers, in case they are 64-bit values.
1276 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1277 test 2 if it fails.
1279 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1280 and a #define of that name to empty if it is not externally set. This is to
1281 allow users of MSVC to set it if necessary.
1283 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1284 the convenience functions in the pcre_get.c source file.
1286 11. An option change at the start of a pattern that had top-level alternatives
1287 could cause overwriting and/or a crash. This command provoked a crash in
1288 some environments:
1290 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1292 This potential security problem was recorded as CVE-2008-2371.
1294 12. For a pattern where the match had to start at the beginning or immediately
1295 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1296 pcre_dfa_exec() could read past the end of the passed subject if there was
1297 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1298 pcretest so that it places the subject at the end of its malloc-ed buffer.
1300 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1301 exec() might read past the end of the data buffer in UTF-8 mode.
1303 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1304 the data contained the byte 0x85 as part of a UTF-8 character within its
1305 first line. This applied both to normal and DFA matching.
1307 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1308 /^[^d]*?$/8 failed to match "abc".
1310 16. Added a missing copyright notice to pcrecpp_internal.h.
1312 17. Make it more clear in the documentation that values returned from
1313 pcre_exec() in ovector are byte offsets, not character counts.
1315 18. Tidied a few places to stop certain compilers from issuing warnings.
1317 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1318 supplied by Stefan Weber. I made a further small update for 7.8 because
1319 there is a change of source arrangements: the pcre_searchfuncs.c module is
1320 replaced by pcre_ucd.c.
1323 Version 7.7 07-May-08
1324 ---------------------
1326 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1327 a string to a long long, pretend we don't even have a long long." This is
1328 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1330 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1331 pre-7.6 versions, which defined a global no_arg variable instead of putting
1332 it in the RE class. (See also #8 below.)
1334 3. Remove a line of dead code, identified by coverity and reported by Nuno
1335 Lopes.
1337 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1339 (1) The include/exclude patterns were being applied to the whole pathnames
1340 of files, instead of just to the final components.
1342 (2) If there was more than one level of directory, the subdirectories were
1343 skipped unless they satisfied the include/exclude conditions. This is
1344 inconsistent with GNU grep (and could even be seen as contrary to the
1345 pcregrep specification - which I improved to make it absolutely clear).
1346 The action now is always to scan all levels of directory, and just
1347 apply the include/exclude patterns to regular files.
1349 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1350 --exclude_dir in the tests to avoid scanning .svn directories.
1352 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1353 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1354 doesn't support NULs in patterns.
1356 7. Added some missing "const"s to declarations of static tables in
1357 pcre_compile.c and pcre_dfa_exec.c.
1359 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1360 caused by fix #2 above. (Subsequently also a second patch to fix the
1361 first patch. And a third patch - this was a messy problem.)
1363 9. Applied Craig's patch to remove the use of push_back().
1365 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1366 matching function regexec().
1368 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1369 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1370 references. PCRE supports relative numbers with this syntax (I don't think
1371 Oniguruma does).
1373 12. Previously, a group with a zero repeat such as (...){0} was completely
1374 omitted from the compiled regex. However, this means that if the group
1375 was called as a subroutine from elsewhere in the pattern, things went wrong
1376 (an internal error was given). Such groups are now left in the compiled
1377 pattern, with a new opcode that causes them to be skipped at execution
1378 time.
1380 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1381 to the way PCRE behaves:
1383 (a) A lone ] character is dis-allowed (Perl treats it as data).
1385 (b) A back reference to an unmatched subpattern matches an empty string
1386 (Perl fails the current match path).
1388 (c) A data ] in a character class must be notated as \] because if the
1389 first data character in a class is ], it defines an empty class. (In
1390 Perl it is not possible to have an empty class.) The empty class []
1391 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1392 The negative empty class [^] matches any one character, independently
1393 of the DOTALL setting.
1395 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1396 non-existent subpattern following a character class starting with ']' and
1397 containing () gave an internal compiling error instead of "reference to
1398 non-existent subpattern". Fortunately, when the pattern did exist, the
1399 compiled code was correct. (When scanning forwards to check for the
1400 existencd of the subpattern, it was treating the data ']' as terminating
1401 the class, so got the count wrong. When actually compiling, the reference
1402 was subsequently set up correctly.)
1404 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1405 it was being rejected as not supported by pcre_dfa_exec(), even though
1406 other assertions are supported. I have made pcre_dfa_exec() support
1407 (*FAIL).
1409 16. The implementation of 13c above involved the invention of a new opcode,
1410 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1411 cannot be changed at match time, I realized I could make a small
1412 improvement to matching performance by compiling OP_ALLANY instead of
1413 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1414 on the OP_ANY path.
1416 17. Compiling pcretest on Windows with readline support failed without the
1417 following two fixes: (1) Make the unistd.h include conditional on
1418 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1420 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1421 ncurses library to be included for pcretest when ReadLine support is
1422 requested, but also to allow for it to be overridden. This patch came from
1423 Daniel Bergström.
1425 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1426 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1427 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1428 spotting this.
1431 Version 7.6 28-Jan-08
1432 ---------------------
1434 1. A character class containing a very large number of characters with
1435 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1436 overflow.
1438 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1439 HAVE_LONG_LONG is not defined.
1441 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1442 bring them up to date and include new features. This patch includes:
1444 - Fixed PH's badly added libz and libbz2 support.
1445 - Fixed a problem with static linking.
1446 - Added pcredemo. [But later removed - see 7 below.]
1447 - Fixed dftables problem and added an option.
1448 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1450 - Added readline support for pcretest.
1451 - Added an listing of the option settings after cmake has run.
1453 4. A user submitted a patch to Makefile that makes it easy to create
1454 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1455 Makefile.am that cause it to include this special target, without
1456 affecting anything else. Note that the same mingw target plus all
1457 the other distribution libraries and programs are now supported
1458 when configuring with CMake (see 6 below) instead of with
1459 Configure/Make.
1461 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1462 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1463 exported in the Windows port". It has not yet been confirmed that the patch
1464 solves the problem, but it does no harm.
1466 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1467 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1468 with CMake, and also correct the comment about stack recursion.
1470 7. Remove the automatic building of pcredemo from the ./configure system and
1471 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1472 of a program that users should build themselves after PCRE is installed, so
1473 building it automatically is not really right. What is more, it gave
1474 trouble in some build environments.
1476 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1479 Version 7.5 10-Jan-08
1480 ---------------------
1482 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1483 values in parens when parsing an RE using the C++ wrapper."
1485 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1486 Characters greater than 255 were excluded from the class instead of being
1487 included.
1489 3. The same bug as (2) above applied to negated POSIX classes such as
1490 [:^space:].
1492 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1493 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1494 I have changed it.
1496 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1497 first named subpattern) and a construct such as (?&a) would reference the
1498 first named subpattern whose name started with "a" (in other words, the
1499 length check was missing). Both these problems are fixed. "Subpattern name
1500 expected" is now given for (?&) (a zero-length name), and this patch also
1501 makes it give the same error for \k'' (previously it complained that that
1502 was a reference to a non-existent subpattern).
1504 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1505 this is right because (?- can be followed by option settings as well as by
1506 digits. I have, however, made the messages clearer.
1508 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1509 than the number used in the conditional) now cause a compile-time error.
1510 This is actually not compatible with Perl, which accepts such patterns, but
1511 treats the conditional as always being FALSE (as PCRE used to), but it
1512 seems to me that giving a diagnostic is better.
1514 8. Change "alphameric" to the more common word "alphanumeric" in comments
1515 and messages.
1517 9. Fix two occurrences of "backslash" in comments that should have been
1518 "backspace".
1520 10. Remove two redundant lines of code that can never be obeyed (their function
1521 was moved elsewhere).
1523 11. The program that makes PCRE's Unicode character property table had a bug
1524 which caused it to generate incorrect table entries for sequences of
1525 characters that have the same character type, but are in different scripts.
1526 It amalgamated them into a single range, with the script of the first of
1527 them. In other words, some characters were in the wrong script. There were
1528 thirteen such cases, affecting characters in the following ranges:
1530 U+002b0 - U+002c1
1531 U+0060c - U+0060d
1532 U+0061e - U+00612
1533 U+0064b - U+0065e
1534 U+0074d - U+0076d
1535 U+01800 - U+01805
1536 U+01d00 - U+01d77
1537 U+01d9b - U+01dbf
1538 U+0200b - U+0200f
1539 U+030fc - U+030fe
1540 U+03260 - U+0327f
1541 U+0fb46 - U+0fbb1
1542 U+10450 - U+1049d
1544 12. The -o option (show only the matching part of a line) for pcregrep was not
1545 compatible with GNU grep in that, if there was more than one match in a
1546 line, it showed only the first of them. It now behaves in the same way as
1547 GNU grep.
1549 13. If the -o and -v options were combined for pcregrep, it printed a blank
1550 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1551 does the same. The return code can be used to tell if there were any
1552 non-matching lines.
1554 14. Added --file-offsets and --line-offsets to pcregrep.
1556 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1557 infinitely looping recursion. The bug was that positive lookaheads were not
1558 being skipped when checking for a possible empty match (negative lookaheads
1559 and both kinds of lookbehind were skipped).
1561 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1562 inclusion of <windows.h> to before rather than after the definition of
1563 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1565 17. Specifying a possessive quantifier with a specific limit for a Unicode
1566 character property caused pcre_compile() to compile bad code, which led at
1567 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1568 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1569 caused the error; without that there was no problem.
1571 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1573 19. Added --enable-pcretest-libreadline.
1575 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1576 RE::GlobalReplace(). As a result, the number of replacements returned was
1577 double what it should be. I removed one of the increments, but Craig sent a
1578 later patch that removed the other one (the right fix) and added unit tests
1579 that check the return values (which was not done before).
1581 21. Several CMake things:
1583 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1584 the names libpcre and libpcreposix, not just pcre and pcreposix.
1586 (2) The above change means that pcretest and pcregrep are now correctly
1587 linked with the newly-built libraries, not previously installed ones.
1591 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1592 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1593 UTF-8 newline character). The key issue is that the pattern starts .*;
1594 this means that the match must be either at the beginning, or after a
1595 newline. The bug was in the code for advancing after a failed match and
1596 checking that the new position followed a newline. It was not taking
1597 account of UTF-8 characters correctly.
1599 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1600 character classes. PCRE was not treating the sequence [:...:] as a
1601 character class unless the ... were all letters. Perl, however, seems to
1602 allow any characters between [: and :], though of course it rejects as
1603 unknown any "names" that contain non-letters, because all the known class
1604 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1605 for example, whereas PCRE did not - it did not recognize a POSIX character
1606 class. This seemed a bit dangerous, so the code has been changed to be
1607 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1608 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1609 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1610 Perl does, and where it didn't before.
1612 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1613 Windows environments %n is disabled by default.
1616 Version 7.4 21-Sep-07
1617 ---------------------
1619 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1620 means that a class such as [\s] counted as "explicit reference to CR or
1621 LF". That isn't really right - the whole point of the change was to try to
1622 help when there was an actual mention of one of the two characters. So now
1623 the change happens only if \r or \n (or a literal CR or LF) character is
1624 encountered.
1626 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1627 of both had grown to the point where there were only 3 bits left.
1628 Fortunately, there was spare space in the data structure, and so I have
1629 moved the internal flags into a new 16-bit field to free up more option
1630 bits.
1632 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1633 but did not set the internal JCHANGED flag - either of these is enough to
1634 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1635 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1636 start sets both bits.
1638 4. Added options (at build time, compile time, exec time) to change \R from
1639 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1641 5. doc/pcresyntax.html was missing from the distribution.
1643 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1644 compatibility, even though it is no longer used.
1646 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1647 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1648 windows.h file is present (where different names are used). [This was
1649 reversed later after testing - see 16 below.]
1651 8. Changed all #include <config.h> to #include "config.h". There were also
1652 some further <pcre.h> cases that I changed to "pcre.h".
1654 9. When pcregrep was used with the --colour option, it missed the line ending
1655 sequence off the lines that it output.
1657 10. It was pointed out to me that arrays of string pointers cause lots of
1658 relocations when a shared library is dynamically loaded. A technique of
1659 using a single long string with a table of offsets can drastically reduce
1660 these. I have refactored PCRE in four places to do this. The result is
1661 dramatic:
1663 Originally: 290
1664 After changing UCP table: 187
1665 After changing error message table: 43
1666 After changing table of "verbs" 36
1667 After changing table of Posix names 22
1669 Thanks to the folks working on Gregex for glib for this insight.
1671 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1672 unicode-properties was also set.
1674 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1676 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1677 checked only for CRLF.
1679 14. Added casts to pcretest.c to avoid compiler warnings.
1681 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1683 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1684 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1685 entirely. This removes changes made in 7 above.
1687 17. The CMake files have been updated, and there is now more information about
1688 building with CMake in the NON-UNIX-USE document.
1691 Version 7.3 28-Aug-07
1692 ---------------------
1694 1. In the rejigging of the build system that eventually resulted in 7.1, the
1695 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1696 brackets there is not right, since it causes compilers to look for an
1697 installed pcre.h, not the version that is in the source that is being
1698 compiled (which of course may be different). I have changed it back to:
1700 #include "pcre.h"
1702 I have a vague recollection that the change was concerned with compiling in
1703 different directories, but in the new build system, that is taken care of
1704 by the VPATH setting the Makefile.
1706 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1707 when the subject happened to end in the byte 0x85 (e.g. if the last
1708 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1709 characters but of course it shouldn't be taken as a newline when it is part
1710 of another character. The bug was that, for an unlimited repeat of . in
1711 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1712 characters when looking for a newline.
1714 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1716 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1717 in debug output.
1719 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1720 long printing in the pcrecpp unittest when running under MinGW.
1722 6. ESC_K was left out of the EBCDIC table.
1724 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1725 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1726 limit also applies to "virtual nesting" when a pattern is recursive, and in
1727 this case 1000 isn't so big. I have been able to remove this limit at the
1728 expense of backing off one optimization in certain circumstances. Normally,
1729 when pcre_exec() would call its internal match() function recursively and
1730 immediately return the result unconditionally, it uses a "tail recursion"
1731 feature to save stack. However, when a subpattern that can match an empty
1732 string has an unlimited repetition quantifier, it no longer makes this
1733 optimization. That gives it a stack frame in which to save the data for
1734 checking that an empty string has been matched. Previously this was taken
1735 from the 1000-entry workspace that had been reserved. So now there is no
1736 explicit limit, but more stack is used.
1738 8. Applied Daniel's patches to solve problems with the import/export magic
1739 syntax that is required for Windows, and which was going wrong for the
1740 pcreposix and pcrecpp parts of the library. These were overlooked when this
1741 problem was solved for the main library.
1743 9. There were some crude static tests to avoid integer overflow when computing
1744 the size of patterns that contain repeated groups with explicit upper
1745 limits. As the maximum quantifier is 65535, the maximum group length was
1746 set at 30,000 so that the product of these two numbers did not overflow a
1747 32-bit integer. However, it turns out that people want to use groups that
1748 are longer than 30,000 bytes (though not repeat them that many times).
1749 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1750 made it possible to implement the integer overflow checks in a much more
1751 dynamic way, which I have now done. The artificial limitation on group
1752 length has been removed - we now have only the limit on the total length of
1753 the compiled pattern, which depends on the LINK_SIZE setting.
1755 10. Fixed a bug in the documentation for get/copy named substring when
1756 duplicate names are permitted. If none of the named substrings are set, the
1757 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1758 empty string.
1760 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1761 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1762 because the ] is interpreted as the first data character and the
1763 terminating ] is not found. PCRE has been made compatible with Perl in this
1764 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1765 cause memory overwriting.
1767 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1768 string has been matched (to stop an infinite loop). It was not recognizing
1769 a conditional subpattern that could match an empty string if that
1770 subpattern was within another subpattern. For example, it looped when
1771 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1772 condition was not nested. This bug has been fixed.
1774 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1775 past the start of the subject in the presence of bytes with the top bit
1776 set, for example "\x8aBCD".
1778 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1779 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1781 14. Optimized (?!) to (*FAIL).
1783 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1784 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1785 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1786 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1787 does: it's just the validity check that is more restrictive.
1789 16. Inserted checks for integer overflows during escape sequence (backslash)
1790 processing, and also fixed erroneous offset values for syntax errors during
1791 backslash processing.
1793 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1794 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1796 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1797 caused an overrun.
1799 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1800 something other than just ASCII characters) inside a group that had an
1801 unlimited repeat caused a loop at compile time (while checking to see
1802 whether the group could match an empty string).
1804 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1805 [\P{Any}] did so. (Error in the code for printing property names.)
1807 21. An orphan \E inside a character class could cause a crash.
1809 22. A repeated capturing bracket such as (A)? could cause a wild memory
1810 reference during compilation.
1812 23. There are several functions in pcre_compile() that scan along a compiled
1813 expression for various reasons (e.g. to see if it's fixed length for look
1814 behind). There were bugs in these functions when a repeated \p or \P was
1815 present in the pattern. These operators have additional parameters compared
1816 with \d, etc, and these were not being taken into account when moving along
1817 the compiled data. Specifically:
1819 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1820 length.
1822 (b) An item such as \pL+ within a repeated group could cause crashes or
1823 loops.
1825 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1826 "reference to non-existent subpattern" error.
1828 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1830 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1831 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1833 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1835 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1836 character were causing crashes (broken optimization).
1838 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1839 \p or \P) caused a compile-time loop.
1841 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1842 break. For example, the unstudied pattern [\r\n]A does not match the string
1843 "\r\nA" because change 7.0/46 below moves the current point on by two
1844 characters after failing to match at the start. However, the pattern \nA
1845 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1846 the same is true. There doesn't seem any very clean way out of this, but
1847 what I have chosen to do makes the common cases work: PCRE now takes note
1848 of whether there can be an explicit match for \r or \n anywhere in the
1849 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1850 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1851 pattern has explicit CR or LF references.
1853 29. Added (*CR) etc for changing newline setting at start of pattern.
1856 Version 7.2 19-Jun-07
1857 ---------------------
1859 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1860 which is apparently normally available under Windows.
1862 2. Re-jig the pcregrep tests with different newline settings in an attempt
1863 to make them independent of the local environment's newline setting.
1865 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1867 4. Some of the "internals" tests were previously cut out when the link size
1868 was not 2, because the output contained actual offsets. The recent new
1869 "Z" feature of pcretest means that these can be cut out, making the tests
1870 usable with all link sizes.
1872 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1873 stack recursion. This gives a massive performance boost under BSD, but just
1874 a small improvement under Linux. However, it saves one field in the frame
1875 in all cases.
1877 6. Added more features from the forthcoming Perl 5.10:
1879 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1880 recursion call. It refers to the nth most recently opened parentheses.
1882 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1883 to be opened parentheses.
1885 (c) Conditions that refer to capturing parentheses can be specified
1886 relatively, for example, (?(-2)... or (?(+3)...
1888 (d) \K resets the start of the current match so that everything before
1889 is not part of it.
1891 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1893 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1894 reference syntax.
1896 (g) (?| introduces a group in which the numbering of parentheses in each
1897 alternative starts with the same number.
1899 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1901 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1904 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1905 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1906 for detecting groups that can match an empty string.
1908 9. A pattern with a very large number of alternatives (more than several
1909 hundred) was running out of internal workspace during the pre-compile
1910 phase, where pcre_compile() figures out how much memory will be needed. A
1911 bit of new cunning has reduced the workspace needed for groups with
1912 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1913 workspace instead of running out of the 4096 that are available.
1915 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1917 11. Applied patch from Google to remove an optimization that didn't quite work.
1918 The report of the bug said:
1920 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1921 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1922 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1924 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1925 it matched the wrong number of bytes.
1928 Version 7.1 24-Apr-07
1929 ---------------------
1931 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1932 that is more "standard", making use of automake and other Autotools. There
1933 is some re-arrangement of the files and adjustment of comments consequent
1934 on this.
1936 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1937 for recursive directory scanning broke on some systems because the files
1938 are not scanned in any specific order and on different systems the order
1939 was different. A call to "sort" has been inserted into RunGrepTest for the
1940 approprate test as a short-term fix. In the longer term there may be an
1941 alternative.
1943 3. I had an email from Eric Raymond about problems translating some of PCRE's
1944 man pages to HTML (despite the fact that I distribute HTML pages, some
1945 people do their own conversions for various reasons). The problems
1946 concerned the use of low-level troff macros .br and .in. I have therefore
1947 removed all such uses from the man pages (some were redundant, some could
1948 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1949 HTML has been updated to handle .nf/.fi and to complain if it encounters
1950 .br or .in.
1952 4. Updated comments in configure.ac that get placed in config.h.in and also
1953 arranged for config.h to be included in the distribution, with the name
1954 config.h.generic, for the benefit of those who have to compile without
1955 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1957 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1958 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1959 makevp.bat for the current PCRE, using the additional files
1960 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1962 6. A Windows user reported a minor discrepancy with test 2, which turned out
1963 to be caused by a trailing space on an input line that had got lost in his
1964 copy. The trailing space was an accident, so I've just removed it.
1966 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1967 that is needed.
1969 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1970 as "const" (a) because they are and (b) because it helps the PHP
1971 maintainers who have recently made a script to detect big data structures
1972 in the php code that should be moved to the .rodata section. I remembered
1973 to update Builducptable as well, so it won't revert if ucptable.h is ever
1974 re-created.
1976 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1977 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1978 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1979 support is not required. This saves 1.5-2K of code, which is important in
1980 some applications.
1982 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1983 so as not to refer to the tables, even though these functions will never be
1984 called when UTF-8 support is disabled. Otherwise there are problems with a
1985 shared library.
1987 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1989 (a) It was defining its arguments as char * instead of void *.
1991 (b) It was assuming that all moves were upwards in memory; this was true
1992 a long time ago when I wrote it, but is no longer the case.
1994 The emulated memove() is provided for those environments that have neither
1995 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1996 is clearly not the case, as these two bugs were recently reported.
1998 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1999 and Detrail to create the HTML documentation, the .txt form of the man
2000 pages, and it removes trailing spaces from listed files. It also creates
2001 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2002 case, it wraps all the #defines with #ifndefs. This script should be run
2003 before "make dist".
2005 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2006 with Unicode property support.
2008 (a) For a maximizing quantifier, if the two different cases of the
2009 character were of different lengths in their UTF-8 codings (there are
2010 some cases like this - I found 11), and the matching function had to
2011 back up over a mixture of the two cases, it incorrectly assumed they
2012 were both the same length.
2014 (b) When PCRE was configured to use the heap rather than the stack for
2015 recursion during matching, it was not correctly preserving the data for
2016 the other case of a UTF-8 character when checking ahead for a match
2017 while processing a minimizing repeat. If the check also involved
2018 matching a wide character, but failed, corruption could cause an
2019 erroneous result when trying to check for a repeat of the original
2020 character.
2022 13. Some tidying changes to the testing mechanism:
2024 (a) The RunTest script now detects the internal link size and whether there
2025 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2026 values substituted by "configure". (The RunGrepTest script already did
2027 this for UTF-8.) The configure.ac script no longer substitutes the
2028 relevant variables.
2030 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2031 with length and offset values. This means that the output is different
2032 for different internal link sizes. Test 2 is skipped for link sizes
2033 other than 2 because of this, bypassing the problem. Unfortunately,
2034 there was also a test in test 3 (the locale tests) that used /B and
2035 failed for link sizes other than 2. Rather than cut the whole test out,
2036 I have added a new /Z option to pcretest that replaces the length and
2037 offset values with spaces. This is now used to make test 3 independent
2038 of link size. (Test 2 will be tidied up later.)
2040 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2041 segmentation fault instead of returning the appropriate error message.
2043 15. In multiline mode when the newline sequence was set to "any", the pattern
2044 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2045 This doesn't seem right; it now treats the CRLF combination as the line
2046 ending, and so does not match in that case. It's only a pattern such as ^$
2047 that would hit this one: something like ^ABC$ would have failed after \r
2048 and then tried again after \r\n.
2050 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2051 in an attempt to make files that differ only in their line terminators
2052 compare equal. This works on Linux.
2054 17. Under certain error circumstances pcregrep might try to free random memory
2055 as it exited. This is now fixed, thanks to valgrind.
2057 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2058 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2059 was because its rules for how to advance for /g after matching an empty
2060 string at the end of a line did not allow for this case. They now check for
2061 it specially.
2063 20. pcretest is supposed to handle patterns and data of any length, by
2064 extending its buffers when necessary. It was getting this wrong when the
2065 buffer for a data line had to be extended.
2067 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2068 CRLF as a newline sequence.
2070 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2071 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2072 I have nevertheless tidied it up.
2074 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2076 24. Added a man page for pcre-config.
2079 Version 7.0 19-Dec-06
2080 ---------------------
2082 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2083 moving to gcc 4.1.1.
2085 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2086 sys/time.h, which is documented as needed for this function. It doesn't
2087 seem to matter on Linux, but it showed up on some releases of OS X.
2089 3. It seems that there are systems where bytes whose values are greater than
2090 127 match isprint() in the "C" locale. The "C" locale should be the
2091 default when a C program starts up. In most systems, only ASCII printing
2092 characters match isprint(). This difference caused the output from pcretest
2093 to vary, making some of the tests fail. I have changed pcretest so that:
2095 (a) When it is outputting text in the compiled version of a pattern, bytes
2096 other than 32-126 are always shown as hex escapes.
2098 (b) When it is outputting text that is a matched part of a subject string,
2099 it does the same, unless a different locale has been set for the match
2100 (using the /L modifier). In this case, it uses isprint() to decide.
2102 4. Fixed a major bug that caused incorrect computation of the amount of memory
2103 required for a compiled pattern when options that changed within the
2104 pattern affected the logic of the preliminary scan that determines the
2105 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2106 that the computed length was too small. The symptoms of this bug were
2107 either the PCRE error "internal error: code overflow" from pcre_compile(),
2108 or a glibc crash with a message such as "pcretest: free(): invalid next
2109 size (fast)". Examples of patterns that provoked this bug (shown in
2110 pcretest format) are:
2112 /(?-x: )/x
2113 /(?x)(?-x: \s*#\s*)/
2114 /((?i)[\x{c0}])/8
2115 /(?i:[\x{c0}])/8
2117 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2118 is now done differently.
2120 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2121 wrapper classes; (b) implement a new function in the C++ scanner that is
2122 more efficient than the old way of doing things because it avoids levels of
2123 recursion in the regex matching; (c) add a paragraph to the documentation
2124 for the FullMatch() function.
2126 6. The escape sequence \n was being treated as whatever was defined as
2127 "newline". Not only was this contrary to the documentation, which states
2128 that \n is character 10 (hex 0A), but it also went horribly wrong when
2129 "newline" was defined as CRLF. This has been fixed.
2131 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2132 was being set to -1 for the "end of line" case (supposedly a value that no
2133 character can have). Though this value is never used (the check for end of
2134 line is "zero bytes in current character"), it caused compiler complaints.
2135 I've changed it to 0xffffffff.
2137 8. In pcre_version.c, the version string was being built by a sequence of
2138 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2139 string (as it is for production releases) called a macro with an empty
2140 argument. The C standard says the result of this is undefined. The gcc
2141 compiler treats it as an empty string (which was what was wanted) but it is
2142 reported that Visual C gives an error. The source has been hacked around to
2143 avoid this problem.
2145 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2146 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2147 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2148 of them did).
2150 10. Originally, pcretest opened its input and output without "b"; then I was
2151 told that "b" was needed in some environments, so it was added for release
2152 5.0 to both the input and output. (It makes no difference on Unix-like
2153 systems.) Later I was told that it is wrong for the input on Windows. I've
2154 now abstracted the modes into two macros, to make it easier to fiddle with
2155 them, and removed "b" from the input mode under Windows.
2157 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2159 12. Added -help and --help to pcretest as an official way of being reminded
2160 of the options.
2162 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2163 and pcrecpp.cc because they annoy compilers at high warning levels.
2165 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2167 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2168 scripts are not C programs :-) and which was not noticed because it works
2169 on Linux.
2171 16. pcretest is supposed to handle any length of pattern and data line (as one
2172 line or as a continued sequence of lines) by extending its input buffer if
2173 necessary. This feature was broken for very long pattern lines, leading to
2174 a string of junk being passed to pcre_compile() if the pattern was longer
2175 than about 50K.
2177 17. I have done a major re-factoring of the way pcre_compile() computes the
2178 amount of memory needed for a compiled pattern. Previously, there was code
2179 that made a preliminary scan of the pattern in order to do this. That was
2180 OK when PCRE was new, but as the facilities have expanded, it has become
2181 harder and harder to keep it in step with the real compile phase, and there
2182 have been a number of bugs (see for example, 4 above). I have now found a
2183 cunning way of running the real compile function in a "fake" mode that
2184 enables it to compute how much memory it would need, while actually only
2185 ever using a few hundred bytes of working memory and without too many
2186 tests of the mode. This should make future maintenance and development
2187 easier. A side effect of this work is that the limit of 200 on the nesting
2188 depth of parentheses has been removed (though this was never a serious
2189 limitation, I suspect). However, there is a downside: pcre_compile() now
2190 runs more slowly than before (30% or more, depending on the pattern). I
2191 hope this isn't a big issue. There is no effect on runtime performance.
2193 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2194 newline (only possible for the last line of a file) and it was a
2195 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2197 19. Added additional timing features to pcretest. (1) The -tm option now times
2198 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2199 separate command line item, by a number that specifies the number of
2200 repeats to use when timing. The default is 50000; this gives better
2201 precision, but takes uncomfortably long for very large patterns.
2203 20. Extended pcre_study() to be more clever in cases where a branch of a
2204 subpattern has no definite first character. For example, (a*|b*)[cd] would
2205 previously give no result from pcre_study(). Now it recognizes that the
2206 first character must be a, b, c, or d.
2208 21. There was an incorrect error "recursive call could loop indefinitely" if
2209 a subpattern (or the entire pattern) that was being tested for matching an
2210 empty string contained only one non-empty item after a nested subpattern.
2211 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2212 incorrectly, because the \d was being skipped in the check.
2214 22. The pcretest program now has a new pattern option /B and a command line
2215 option -b, which is equivalent to adding /B to every pattern. This causes
2216 it to show the compiled bytecode, without the additional information that
2217 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2218 is the same as /B/I).
2220 23. A new optimization is now able automatically to treat some sequences such
2221 as a*b as a*+b. More specifically, if something simple (such as a character
2222 or a simple class like \d) has an unlimited quantifier, and is followed by
2223 something that cannot possibly match the quantified thing, the quantifier
2224 is automatically "possessified".
2226 24. A recursive reference to a subpattern whose number was greater than 39
2227 went wrong under certain circumstances in UTF-8 mode. This bug could also
2228 have affected the operation of pcre_study().
2230 25. Realized that a little bit of performance could be had by replacing
2231 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2233 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2235 27. Possessive quantifiers such as a++ were previously implemented by turning
2236 them into atomic groups such as ($>a+). Now they have their own opcodes,
2237 which improves performance. This includes the automatically created ones
2238 from 23 above.
2240 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2241 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2242 the first matched character to be a colon. This applied both to named and
2243 numbered groups.
2245 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2247 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2248 building PCRE on Windows easier, so I have included it in the distribution.
2250 31. There is now a check in pcretest against a ridiculously large number being
2251 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2252 loop, the loop is abandoned.
2254 32. Forward references to subpatterns in conditions such as (?(2)...) where
2255 subpattern 2 is defined later cause pcre_compile() to search forwards in
2256 the pattern for the relevant set of parentheses. This search went wrong
2257 when there were unescaped parentheses in a character class, parentheses
2258 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2260 33. "Subroutine" calls and backreferences were previously restricted to
2261 referencing subpatterns earlier in the regex. This restriction has now
2262 been removed.
2264 34. Added a number of extra features that are going to be in Perl 5.10. On the
2265 whole, these are just syntactic alternatives for features that PCRE had
2266 previously implemented using the Python syntax or my own invention. The
2267 other formats are all retained for compatibility.
2269 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2270 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2271 also .NET compatible.
2273 (b) A recursion or subroutine call to a named group can now be defined as
2274 (?&name) as well as (?P>name).
2276 (c) A backreference to a named group can now be defined as \k<name> or
2277 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2278 5.10, are also .NET compatible.
2280 (d) A conditional reference to a named group can now use the syntax
2281 (?(<name>) or (?('name') as well as (?(name).
2283 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2284 groups (named and numbered) that are never evaluated inline, but can be
2285 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2286 is always false. There may be only one alternative in such a group.
2288 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2289 as the simple (?(R). The condition is true only if the most recent
2290 recursion is that of the given number or name. It does not search out
2291 through the entire recursion stack.
2293 (g) The escape \gN or \g{N} has been added, where N is a positive or
2294 negative number, specifying an absolute or relative reference.
2296 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2297 some "unreachable code" warnings.
2299 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2300 things, this adds five new scripts.
2302 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2303 There were also incompatibilities regarding the handling of \Q..\E inside
2304 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2305 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2307 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2308 matches an empty string, and forcibly breaks the loop. There were bugs in
2309 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2310 against aaaa the result was just "a" rather than "aaaa", for example. Two
2311 separate and independent bugs (that affected different cases) have been
2312 fixed.
2314 39. Refactored the code to abolish the use of different opcodes for small
2315 capturing bracket numbers. This is a tidy that I avoided doing when I
2316 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2317 The new approach is not only tidier, it makes it possible to reduce the
2318 memory needed to fix the previous bug (38).
2320 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2321 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2322 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2323 mode.
2325 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2326 report.
2328 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2329 copy construction and assignment in the C++ wrapper.
2331 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2332 couple of bugs that could have given wrong results in the "--newline=crlf"
2333 case.
2335 44. Added a number of casts and did some reorganization of signed/unsigned int
2336 variables following suggestions from Dair Grant. Also renamed the variable
2337 "this" as "item" because it is a C++ keyword.
2339 45. Arranged for dftables to add
2341 #include "pcre_internal.h"
2343 to pcre_chartables.c because without it, gcc 4.x may remove the array
2344 definition from the final binary if PCRE is built into a static library and
2345 dead code stripping is activated.
2347 46. For an unanchored pattern, if a match attempt fails at the start of a
2348 newline sequence, and the newline setting is CRLF or ANY, and the next two
2349 characters are CRLF, advance by two characters instead of one.
2352 Version 6.7 04-Jul-06
2353 ---------------------
2355 1. In order to handle tests when input lines are enormously long, pcretest has
2356 been re-factored so that it automatically extends its buffers when
2357 necessary. The code is crude, but this _is_ just a test program. The
2358 default size has been increased from 32K to 50K.
2360 2. The code in pcre_study() was using the value of the re argument before
2361 testing it for NULL. (Of course, in any sensible call of the function, it
2362 won't be NULL.)
2364 3. The memmove() emulation function in pcre_internal.h, which is used on
2365 systems that lack both memmove() and bcopy() - that is, hardly ever -
2366 was missing a "static" storage class specifier.
2368 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2369 containing an extended class (one that cannot be represented by a bitmap
2370 because it contains high-valued characters or Unicode property items, e.g.
2371 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2372 pattern, but PCRE should not loop if you do not (it no longer does).
2373 [Detail: two cases were found: (a) a repeated subpattern containing an
2374 extended class; (b) a recursive reference to a subpattern that followed a
2375 previous extended class. It wasn't skipping over the extended class
2376 correctly when UTF-8 mode was not set.]
2378 5. A negated single-character class was not being recognized as fixed-length
2379 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2380 compile error "lookbehind assertion is not fixed length".
2382 6. The RunPerlTest auxiliary script was showing an unexpected difference
2383 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2384 write a Perl script that can interpret lines of an input file either as
2385 byte characters or as UTF-8, which is what "perltest" was being required to
2386 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2387 can't do is switch easily at run time between having the "use utf8;" pragma
2388 or not. In the end, I fudged it by using the RunPerlTest script to insert
2389 "use utf8;" explicitly for the UTF-8 tests.
2391 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2392 the end of the subject string, contrary to the documentation and to what
2393 Perl does. This was true of both matching functions. Now it matches only at
2394 the start of the subject and immediately after *internal* newlines.
2396 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2397 a pointer to an int instead of a pointer to an unsigned long int. This
2398 caused problems on 64-bit systems.
2400 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2401 instance of the 'standard' template library not being so standard".
2403 10. There was no check on the number of named subpatterns nor the maximum
2404 length of a subpattern name. The product of these values is used to compute
2405 the size of the memory block for a compiled pattern. By supplying a very
2406 long subpattern name and a large number of named subpatterns, the size
2407 computation could be caused to overflow. This is now prevented by limiting
2408 the length of names to 32 characters, and the number of named subpatterns
2409 to 10,000.
2411 11. Subpatterns that are repeated with specific counts have to be replicated in
2412 the compiled pattern. The size of memory for this was computed from the
2413 length of the subpattern and the repeat count. The latter is limited to
2414 65535, but there was no limit on the former, meaning that integer overflow
2415 could in principle occur. The compiled length of a repeated subpattern is
2416 now limited to 30,000 bytes in order to prevent this.
2418 12. Added the optional facility to have named substrings with the same name.
2420 13. Added the ability to use a named substring as a condition, using the
2421 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2422 are numbers (not recommended). Forward references are permitted.
2424 14. Added forward references in named backreferences (if you see what I mean).
2426 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2427 pattern could run off the end of the subject. For example, the pattern
2428 "(?s)(.{1,5})"8 did this with the subject "ab".
2430 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2431 PCRE_CASELESS was set when matching characters that were quantified with ?
2432 or *.
2434 17. A character class other than a single negated character that had a minimum
2435 but no maximum quantifier - for example [ab]{6,} - was not handled
2436 correctly by pce_dfa_exec(). It would match only one character.
2438 18. A valid (though odd) pattern that looked like a POSIX character
2439 class but used an invalid character after [ (for example [[,abc,]]) caused
2440 pcre_compile() to give the error "Failed: internal error: code overflow" or
2441 in some cases to crash with a glibc free() error. This could even happen if
2442 the pattern terminated after [[ but there just happened to be a sequence of
2443 letters, a binary zero, and a closing ] in the memory that followed.
2445 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2446 over the years. Originally (before any Unicode support), just the bottom 8
2447 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2448 output from "man perlunicode" includes this:
2450 The regular expression compiler produces polymorphic opcodes. That
2451 is, the pattern adapts to the data and automatically switches to
2452 the Unicode character scheme when presented with Unicode data--or
2453 instead uses a traditional byte scheme when presented with byte
2454 data.
2456 Sadly, a wide octal escape does not cause a switch, and in a string with
2457 no other multibyte characters, these octal escapes are treated as before.
2458 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2459 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2460 Unicode string.
2462 I have not perpetrated such confusion in PCRE. Up till now, it took just
2463 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2464 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2465 translate to the appropriate multibyte character.
2467 29. Applied some refactoring to reduce the number of warnings from Microsoft
2468 and Borland compilers. This has included removing the fudge introduced
2469 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2470 a warning about an unused variable.
2472 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2473 characters since release 4.0, because Perl (from release 5.004) does not.
2474 [Or at least, is documented not to: some releases seem to be in conflict
2475 with the documentation.] However, when a pattern was studied with
2476 pcre_study() and all its branches started with \s, PCRE still included VT
2477 as a possible starting character. Of course, this did no harm; it just
2478 caused an unnecessary match attempt.
2480 22. Removed a now-redundant internal flag bit that recorded the fact that case
2481 dependency changed within the pattern. This was once needed for "required
2482 byte" processing, but is no longer used. This recovers a now-scarce options
2483 bit. Also moved the least significant internal flag bit to the most-
2484 significant bit of the word, which was not previously used (hangover from
2485 the days when it was an int rather than a uint) to free up another bit for
2486 the future.
2488 23. Added support for CRLF line endings as well as CR and LF. As well as the
2489 default being selectable at build time, it can now be changed at runtime
2490 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2491 specify that it is scanning data with non-default line endings.
2493 24. Changed the definition of CXXLINK to make it agree with the definition of
2494 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2496 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2497 recursions. This makes a big different to stack usage for some patterns.
2499 26. If a subpattern containing a named recursion or subroutine reference such
2500 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2501 the space required for the compiled pattern went wrong and gave too small a
2502 value. Depending on the environment, this could lead to "Failed: internal
2503 error: code overflow at offset 49" or "glibc detected double free or
2504 corruption" errors.
2506 27. Applied patches from Google (a) to support the new newline modes and (b) to
2507 advance over multibyte UTF-8 characters in GlobalReplace.
2509 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2510 difference for some implementation of PCRE in some Windows version.
2512 29. Added some extra testing facilities to pcretest:
2514 \q<number> in a data line sets the "match limit" value
2515 \Q<number> in a data line sets the "match recursion limt" value
2516 -S <number> sets the stack size, where <number> is in megabytes
2518 The -S option isn't available for Windows.
2521 Version 6.6 06-Feb-06
2522 ---------------------
2524 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2525 in pcreposix.h. I have copied the definition from pcre.h.
2527 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2528 because pcre.h is no longer a built file.
2530 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2531 not normally included in the compiled code.
2534 Version 6.5 01-Feb-06
2535 ---------------------
2537 1. When using the partial match feature with pcre_dfa_exec(), it was not
2538 anchoring the second and subsequent partial matches at the new starting
2539 point. This could lead to incorrect results. For example, with the pattern
2540 /1234/, partially matching against "123" and then "a4" gave a match.
2542 2. Changes to pcregrep:
2544 (a) All non-match returns from pcre_exec() were being treated as failures
2545 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2546 error message is output. Some extra information is given for the
2548 probably the only errors that are likely to be caused by users (by
2549 specifying a regex that has nested indefinite repeats, for instance).
2550 If there are more than 20 of these errors, pcregrep is abandoned.
2552 (b) A binary zero was treated as data while matching, but terminated the
2553 output line if it was written out. This has been fixed: binary zeroes
2554 are now no different to any other data bytes.
2556 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2557 used to set a locale for matching. The --locale=xxxx long option has
2558 been added (no short equivalent) to specify a locale explicitly on the
2559 pcregrep command, overriding the environment variables.
2561 (d) When -B was used with -n, some line numbers in the output were one less
2562 than they should have been.
2564 (e) Added the -o (--only-matching) option.
2566 (f) If -A or -C was used with -c (count only), some lines of context were
2567 accidentally printed for the final match.
2569 (g) Added the -H (--with-filename) option.
2571 (h) The combination of options -rh failed to suppress file names for files
2572 that were found from directory arguments.
2574 (i) Added the -D (--devices) and -d (--directories) options.
2576 (j) Added the -F (--fixed-strings) option.
2578 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2580 (l) Added the --colo(u)r option.
2582 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2583 is not present by default.
2585 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2586 items such as (?R) or (?1), when the recursion could match a number of
2587 alternatives. If it matched one of the alternatives, but subsequently,
2588 outside the recursion, there was a failure, the code tried to back up into
2589 the recursion. However, because of the way PCRE is implemented, this is not
2590 possible, and the result was an incorrect result from the match.
2592 In order to prevent this happening, the specification of recursion has
2593 been changed so that all such subpatterns are automatically treated as
2594 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2596 4. I had overlooked the fact that, in some locales, there are characters for
2597 which isalpha() is true but neither isupper() nor islower() are true. In
2598 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2599 and ordfeminine) are like this. This affected the treatment of \w and \W
2600 when they appeared in character classes, but not when they appeared outside
2601 a character class. The bit map for "word" characters is now created
2602 separately from the results of isalnum() instead of just taking it from the
2603 upper, lower, and digit maps. (Plus the underscore character, of course.)
2605 5. The above bug also affected the handling of POSIX character classes such as
2606 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2607 permanent tables. Instead, the bit maps for such a class were previously
2608 created as the appropriate unions of the upper, lower, and digit bitmaps.
2609 Now they are created by subtraction from the [[:word:]] class, which has
2610 its own bitmap.
2612 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2613 It is created by subtracting the vertical space characters (\x09, \x0a,
2614 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2615 subtraction was done in the overall bitmap for a character class, meaning
2616 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2617 be recognized. This bug has been fixed.
2619 7. Patches from the folks at Google:
2621 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2622 real life, but is still worth protecting against".
2624 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2625 regular expressions".
2627 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2628 have it.
2630 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2631 "configure" and the latter not, in order to fix a problem somebody had
2632 with compiling the Arg class on HP-UX.
2634 (e) Improve the error-handling of the C++ wrapper a little bit.
2636 (f) New tests for checking recursion limiting.
2638 8. The pcre_memmove() function, which is used only if the environment does not
2639 have a standard memmove() function (and is therefore rarely compiled),
2640 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2641 returning a result (though PCRE never actually uses the result).
2643 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2644 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2645 returned instead of calling malloc() with an overflowing number that would
2646 most likely cause subsequent chaos.
2648 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2650 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2651 with this option is matched, the nmatch and pmatch options of regexec() are
2652 ignored.
2654 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2655 provided in case anyone wants to the the POSIX interface with UTF-8
2656 strings.
2658 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2659 C++ linking (needed for some HP-UX environments).
2661 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2662 (unused parameter) and in the pcre_printint() function (omitted "default"
2663 switch label when the default is to do nothing).
2665 15. Added some code to make it possible, when PCRE is compiled as a C++
2666 library, to replace subject pointers for pcre_exec() with a smart pointer
2667 class, thus making it possible to process discontinuous strings.
2669 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2670 much the same function. They were added by different people who were trying
2671 to make PCRE easy to compile on non-Unix systems. It has been suggested
2672 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2673 for compiling on Windows systems. I have therefore replaced it with
2674 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2675 defaults to "extern" for C or "extern C" for C++, which works fine on
2676 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2677 SCOPE with something explicit in config.h. In addition:
2679 (a) pcreposix.h still had just "extern" instead of either of these macros;
2680 I have replaced it with PCRE_DATA_SCOPE.
2682 (b) Functions such as _pcre_xclass(), which are internal to the library,
2683 but external in the C sense, all had PCRE_EXPORT in their definitions.
2684 This is apparently wrong for the Windows case, so I have removed it.
2685 (It makes no difference on Unix-like systems.)
2687 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2688 of recursive calls to match(). This is different to MATCH_LIMIT because
2689 that limits the total number of calls to match(), not all of which increase
2690 the depth of recursion. Limiting the recursion depth limits the amount of
2691 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2692 when PCRE is compiled, and changed at run time. A patch from Google adds
2693 this functionality to the C++ interface.
2695 18. Changes to the handling of Unicode character properties:
2697 (a) Updated the table to Unicode 4.1.0.
2699 (b) Recognize characters that are not in the table as "Cn" (undefined).
2701 (c) I revised the way the table is implemented to a much improved format
2702 which includes recognition of ranges. It now supports the ranges that
2703 are defined in UnicodeData.txt, and it also amalgamates other
2704 characters into ranges. This has reduced the number of entries in the
2705 table from around 16,000 to around 3,000, thus reducing its size
2706 considerably. I realized I did not need to use a tree structure after
2707 all - a binary chop search is just as efficient. Having reduced the
2708 number of entries, I extended their size from 6 bytes to 8 bytes to
2709 allow for more data.
2711 (d) Added support for Unicode script names via properties such as \p{Han}.
2713 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2714 matching that character.
2716 20. When matching a repeated Unicode property with a minimum greater than zero,
2717 (for example \pL{2,}), PCRE could look past the end of the subject if it
2718 reached it while seeking the minimum number of characters. This could
2719 happen only if some of the characters were more than one byte long, because
2720 there is a check for at least the minimum number of bytes.
2722 21. Refactored the implementation of \p and \P so as to be more general, to
2723 allow for more different types of property in future. This has changed the
2724 compiled form incompatibly. Anybody with saved compiled patterns that use
2725 \p or \P will have to recompile them.
2727 22. Added "Any" and "L&" to the supported property types.
2729 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2730 but give a compile time error if the value is greater than 0xff.
2732 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2733 accidentally not being installed or uninstalled.
2735 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2736 made were to insert the current release number. This seemed silly, because
2737 it made things harder for people building PCRE on systems that don't run
2738 "configure". I have turned pcre.h into a distributed file, no longer built
2739 by "configure", with the version identification directly included. There is
2740 no longer a pcre.h.in file.
2742 However, this change necessitated a change to the pcre-config script as
2743 well. It is built from pcre-config.in, and one of the substitutions was the
2744 release number. I have updated configure.ac so that ./configure now finds
2745 the release number by grepping pcre.h.
2747 26. Added the ability to run the tests under valgrind.
2750 Version 6.4 05-Sep-05
2751 ---------------------
2753 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2754 "--" to be printed when multiple files were scanned, even when none of the
2755 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2756 consider it to be a bug, and have restored the previous behaviour.
2758 2. A couple of code tidies to get rid of compiler warnings.
2760 3. The pcretest program used to cheat by referring to symbols in the library
2761 whose names begin with _pcre_. These are internal symbols that are not
2762 really supposed to be visible externally, and in some environments it is
2763 possible to suppress them. The cheating is now confined to including
2764 certain files from the library's source, which is a bit cleaner.
2766 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2767 file's purpose clearer.
2769 5. Reorganized pcre_ucp_findchar().
2772 Version 6.3 15-Aug-05
2773 ---------------------
2775 1. The file libpcre.pc.in did not have general read permission in the tarball.
2777 2. There were some problems when building without C++ support:
2779 (a) If C++ support was not built, "make install" and "make test" still
2780 tried to test it.
2782 (b) There were problems when the value of CXX was explicitly set. Some
2783 changes have been made to try to fix these, and ...
2785 (c) --disable-cpp can now be used to explicitly disable C++ support.
2787 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2788 backslash in a target when C++ was disabled. This confuses some
2789 versions of "make", apparently. Using an intermediate variable solves
2790 this. (Same for CPP_LOBJ.)
2792 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2793 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2794 necessary on certain architectures.
2796 4. Added a setting of -export-symbols-regex to the link command to remove
2797 those symbols that are exported in the C sense, but actually are local
2798 within the library, and not documented. Their names all begin with
2799 "_pcre_". This is not a perfect job, because (a) we have to except some
2800 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2801 available (and never for static libraries). I have made a note to try to
2802 find a way round (a) in the future.
2805 Version 6.2 01-Aug-05
2806 ---------------------
2808 1. There was no test for integer overflow of quantifier values. A construction
2809 such as {1111111111111111} would give undefined results. What is worse, if
2810 a minimum quantifier for a parenthesized subpattern overflowed and became
2811 negative, the calculation of the memory size went wrong. This could have
2812 led to memory overwriting.
2814 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2816 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2817 operating environments where this matters.
2819 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2820 PCRE options from within the C++ wrapper.
2822 5. Named capturing subpatterns were not being correctly counted when a pattern
2823 was compiled. This caused two problems: (a) If there were more than 100
2824 such subpatterns, the calculation of the memory needed for the whole
2825 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2826 back references of the form \12, where the number was greater than 9, were
2827 not recognized as back references, even though there were sufficient
2828 previous subpatterns.
2830 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2831 versions of gcc, e.g. 2.95.4.
2834 Version 6.1 21-Jun-05
2835 ---------------------
2837 1. There was one reference to the variable "posix" in pcretest.c that was not
2838 surrounded by "#if !defined NOPOSIX".
2840 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2841 the cross-check on the old pcre_info() function, for the benefit of the
2842 cut-down version of PCRE that is currently imported into Exim.
2844 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2845 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2846 patterns that start (?i) but I don't think that matters. The use of (?i) is
2847 just an example; this all applies to the other options as well.
2849 4. Since libtool seems to echo the compile commands it is issuing, the output
2850 from "make" can be reduced a bit by putting "@" in front of each libtool
2851 compile command.
2853 5. Patch from the folks at Google for configure.in to be a bit more thorough
2854 in checking for a suitable C++ installation before trying to compile the
2855 C++ stuff. This should fix a reported problem when a compiler was present,
2856 but no suitable headers.
2858 6. The man pages all had just "PCRE" as their title. I have changed them to
2859 be the relevant file name. I have also arranged that these names are
2860 retained in the file doc/pcre.txt, which is a concatenation in text format
2861 of all the man pages except the little individual ones for each function.
2863 7. The NON-UNIX-USE file had not been updated for the different set of source
2864 files that come with release 6. I also added a few comments about the C++
2865 wrapper.
2868 Version 6.0 07-Jun-05
2869 ---------------------
2871 1. Some minor internal re-organization to help with my DFA experiments.
2873 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2874 didn't matter for the library itself when fully configured, but did matter
2875 when compiling without UCP support, or within Exim, where the ucp files are
2876 not imported.
2878 3. Refactoring of the library code to split up the various functions into
2879 different source modules. The addition of the new DFA matching code (see
2880 below) to a single monolithic source would have made it really too
2881 unwieldy, quite apart from causing all the code to be include in a
2882 statically linked application, when only some functions are used. This is
2883 relevant even without the DFA addition now that patterns can be compiled in
2884 one application and matched in another.
2886 The downside of splitting up is that there have to be some external
2887 functions and data tables that are used internally in different modules of
2888 the library but which are not part of the API. These have all had their
2889 names changed to start with "_pcre_" so that they are unlikely to clash
2890 with other external names.
2892 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2893 a different (DFA) algorithm. Although it is slower than the original
2894 function, it does have some advantages for certain types of matching
2895 problem.
2897 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2898 including restarting after a partial match.
2900 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2901 defined when compiling for Windows was sent to me. I have put it into the
2902 code, though I have no means of testing or verifying it.
2904 7. Added the pcre_refcount() auxiliary function.
2906 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2907 match before or at the first newline in the subject string. In pcretest,
2908 the /f option on a pattern can be used to set this.
2910 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2911 would behave wrongly. This has been present in PCRE since release 4.0.
2913 10. A number of changes to the pcregrep command:
2915 (a) Refactored how -x works; insert ^(...)$ instead of setting
2916 PCRE_ANCHORED and checking the length, in preparation for adding
2917 something similar for -w.
2919 (b) Added the -w (match as a word) option.
2921 (c) Refactored the way lines are read and buffered so as to have more
2922 than one at a time available.
2924 (d) Implemented a pcregrep test script.
2926 (e) Added the -M (multiline match) option. This allows patterns to match
2927 over several lines of the subject. The buffering ensures that at least
2928 8K, or the rest of the document (whichever is the shorter) is available
2929 for matching (and similarly the previous 8K for lookbehind assertions).
2931 (f) Changed the --help output so that it now says
2933 -w, --word-regex(p)
2935 instead of two lines, one with "regex" and the other with "regexp"
2936 because that confused at least one person since the short forms are the
2937 same. (This required a bit of code, as the output is generated
2938 automatically from a table. It wasn't just a text change.)
2940 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2941 option but starts with a hyphen. Could be a pattern or a path name
2942 starting with a hyphen, for instance.
2944 (h) "-" can be given as a file name to represent stdin.
2946 (i) When file names are being printed, "(standard input)" is used for
2947 the standard input, for compatibility with GNU grep. Previously
2948 "<stdin>" was used.
2950 (j) The option --label=xxx can be used to supply a name to be used for
2951 stdin when file names are being printed. There is no short form.
2953 (k) Re-factored the options decoding logic because we are going to add
2954 two more options that take data. Such options can now be given in four
2955 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2957 (l) Added the -A, -B, and -C options for requesting that lines of context
2958 around matches be printed.
2960 (m) Added the -L option to print the names of files that do not contain
2961 any matching lines, that is, the complement of -l.
2963 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2964 continue to scan other files.
2966 (o) The -s option was incorrectly implemented. For compatibility with other
2967 greps, it now suppresses the error message for a non-existent or non-
2968 accessible file (but not the return code). There is a new option called
2969 -q that suppresses the output of matching lines, which was what -s was
2970 previously doing.
2972 (p) Added --include and --exclude options to specify files for inclusion
2973 and exclusion when recursing.
2975 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2976 Hopefully, it now does.
2978 12. Missing cast in pcre_study().
2980 13. Added an "uninstall" target to the makefile.
2982 14. Replaced "extern" in the function prototypes in Makefile.in with
2983 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2984 world, but is set differently for Windows.
2986 15. Added a second compiling function called pcre_compile2(). The only
2987 difference is that it has an extra argument, which is a pointer to an
2988 integer error code. When there is a compile-time failure, this is set
2989 non-zero, in addition to the error test pointer being set to point to an
2990 error message. The new argument may be NULL if no error number is required
2991 (but then you may as well call pcre_compile(), which is now just a
2992 wrapper). This facility is provided because some applications need a
2993 numeric error indication, but it has also enabled me to tidy up the way
2994 compile-time errors are handled in the POSIX wrapper.
2996 16. Added VPATH=.libs to the makefile; this should help when building with one
2997 prefix path and installing with another. (Or so I'm told by someone who
2998 knows more about this stuff than I do.)
3000 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3001 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3002 match everything, including newlines. This is not POSIX-compatible, but
3003 somebody wanted the feature. From pcretest it can be activated by using
3004 both the P and the s flags.
3006 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3008 19. libpcre.pc was being incorrectly installed as executable.
3010 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3011 it now also looks for '\r' so that it will work unmodified on Windows.
3013 21. Added Google's contributed C++ wrapper to the distribution.
3015 22. Added some untidy missing memory free() calls in pcretest, to keep
3016 Electric Fence happy when testing.
3020 Version 5.0 13-Sep-04
3021 ---------------------
3023 1. Internal change: literal characters are no longer packed up into items
3024 containing multiple characters in a single byte-string. Each character
3025 is now matched using a separate opcode. However, there may be more than one
3026 byte in the character in UTF-8 mode.
3028 2. The pcre_callout_block structure has two new fields: pattern_position and
3029 next_item_length. These contain the offset in the pattern to the next match
3030 item, and its length, respectively.
3032 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3033 insertion of callouts before each pattern item. Added the /C option to
3034 pcretest to make use of this.
3036 4. On the advice of a Windows user, the lines
3038 #if defined(_WIN32) || defined(WIN32)
3039 _setmode( _fileno( stdout ), 0x8000 );
3040 #endif /* defined(_WIN32) || defined(WIN32) */
3042 have been added to the source of pcretest. This apparently does useful
3043 magic in relation to line terminators.
3045 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3046 for the benefit of those environments where the "b" makes a difference.
3048 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3049 to know about it. I have put a hack into configure.in that adds in code
3050 to set GCC=yes if CC=icc. This seems to end up at a point in the
3051 generated configure script that is early enough to affect the setting of
3052 compiler options, which is what is needed, but I have no means of testing
3053 whether it really works. (The user who reported this had patched the
3054 generated configure script, which of course I cannot do.)
3056 LATER: After change 22 below (new libtool files), the configure script
3057 seems to know about icc (and also ecc). Therefore, I have commented out
3058 this hack in configure.in.
3060 7. Added support for pkg-config (2 patches were sent in).
3062 8. Negated POSIX character classes that used a combination of internal tables
3063 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3064 [[:^ascii]]. Typically, they would match almost any characters. The other
3065 POSIX classes were not broken in this way.
3067 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3068 to find the match, as PCRE was deluded into thinking that the match had to
3069 start at the start point or following a newline. The same bug applied to
3070 patterns with negative forward assertions or any backward assertions
3071 preceding ".*" at the start, unless the pattern required a fixed first
3072 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3074 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3075 starting at the last subject character, bytes beyond the end of the subject
3076 string were read.
3078 11. Renamed the variable "class" as "classbits" to make life easier for C++
3079 users. (Previously there was a macro definition, but it apparently wasn't
3080 enough.)
3082 12. Added the new field "tables" to the extra data so that tables can be passed
3083 in at exec time, or the internal tables can be re-selected. This allows
3084 a compiled regex to be saved and re-used at a later time by a different
3085 program that might have everything at different addresses.
3087 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3088 -R library as well as a -L library.
3090 14. The debugging options of pcretest (-d on the command line or D on a
3091 pattern) showed incorrect output for anything following an extended class
3092 that contained multibyte characters and which was followed by a quantifier.
3094 15. Added optional support for general category Unicode character properties
3095 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3096 support. It adds about 90K to the size of the library. The meanings of the
3097 inbuilt class escapes such as \d and \s have NOT been changed.
3099 16. Updated pcredemo.c to include calls to free() to release the memory for the
3100 compiled pattern.
3102 17. The generated file chartables.c was being created in the source directory
3103 instead of in the building directory. This caused the build to fail if the
3104 source directory was different from the building directory, and was
3105 read-only.
3107 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3108 file. No doubt somebody will tell me if they don't make sense... Also added
3109 Dan Mooney's comments about building on OpenVMS.
3111 19. Added support for partial matching via the PCRE_PARTIAL option for
3112 pcre_exec() and the \P data escape in pcretest.
3114 20. Extended pcretest with 3 new pattern features:
3116 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3117 write the compiled pattern to the file whose name is "rest-of-line".
3118 This is a straight binary dump of the data, with the saved pointer to
3119 the character tables forced to be NULL. The study data, if any, is
3120 written too. After writing, pcretest reads a new pattern.
3122 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3123 compiled pattern from the given file. There must not be any
3124 occurrences of "<" in the file name (pretty unlikely); if there are,
3125 pcretest will instead treat the initial "<" as a pattern delimiter.
3126 After reading in the pattern, pcretest goes on to read data lines as
3127 usual.
3129 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3130 and 16-bit fields in a compiled pattern, to simulate a pattern that
3131 was compiled on a host of opposite endianness.
3133 21. The pcre-exec() function can now cope with patterns that were compiled on
3134 hosts of opposite endianness, with this restriction:
3136 As for any compiled expression that is saved and used later, the tables
3137 pointer field cannot be preserved; the extra_data field in the arguments
3138 to pcre_exec() should be used to pass in a tables address if a value
3139 other than the default internal tables were used at compile time.
3141 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3142 now diagnosed as an error. Previously, most of the time, a negative number
3143 would have been treated as zero, but if in addition "ovector" was passed as
3144 NULL, a crash could occur.
3146 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3147 new versions from the libtool 1.5 distribution (the last one is a copy of
3148 a file called libtool.m4). This seems to have fixed the need to patch
3149 "configure" to support Darwin 1.3 (which I used to do). However, I still
3150 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3151 workstation).
3153 24. Changed the PCRE licence to be the more standard "BSD" licence.
3156 Version 4.5 01-Dec-03
3157 ---------------------
3159 1. There has been some re-arrangement of the code for the match() function so
3160 that it can be compiled in a version that does not call itself recursively.
3161 Instead, it keeps those local variables that need separate instances for
3162 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3163 needs to "recurse". Keeping track of where control must go is done by means
3164 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3165 hide most of the details from the main code, and operates only if
3166 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3167 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3168 operating.
3170 To make it easier for callers to provide specially tailored get/free
3171 functions for this usage, two new functions, pcre_stack_malloc, and
3172 pcre_stack_free, are used. They are always called in strict stacking order,
3173 and the size of block requested is always the same.
3175 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3176 PCRE has been compiled to use the stack or the heap for recursion. The
3177 -C option of pcretest uses this to show which version is compiled.
3179 A new data escape \S, is added to pcretest; it causes the amounts of store
3180 obtained and freed by both kinds of malloc/free at match time to be added
3181 to the output.
3183 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3184 what's available on my current Linux desktop machine.
3186 3. When matching a UTF-8 string, the test for a valid string at the start has
3187 been extended. If start_offset is not zero, PCRE now checks that it points
3188 to a byte that is the start of a UTF-8 character. If not, it returns
3189 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3190 this is necessary because there may be backward assertions in the pattern.
3191 When matching the same subject several times, it may save resources to use
3192 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3194 4. The code for checking the validity of UTF-8 strings has been tightened so
3195 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3196 containing "overlong sequences".
3198 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3199 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3200 should have been "&", but it just so happened that all the cases this let
3201 through by mistake were picked up later in the function.
3203 6. I had used a variable called "isblank" - this is a C99 function, causing
3204 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3206 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3207 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3208 so the version is shown; (c) stop "make test" echoing "./RunTest".
3210 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3212 9. The prototype for memmove() for systems that don't have it was using
3213 size_t, but the inclusion of the header that defines size_t was later. I've
3214 moved the #includes for the C headers earlier to avoid this.
3216 10. Added some adjustments to the code to make it easier to compiler on certain
3217 special systems:
3219 (a) Some "const" qualifiers were missing.
3220 (b) Added the macro EXPORT before all exported functions; by default this
3221 is defined to be empty.
3222 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3223 that it reads its output file name as an argument instead of writing
3224 to the standard output and assuming this can be redirected.
3226 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3227 class containing characters with values greater than 255, PCRE compilation
3228 went into a loop.
3230 12. A recursive reference to a subpattern that was within another subpattern
3231 that had a minimum quantifier of zero caused PCRE to crash. For example,
3232 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3233 recursion. If the recursively-called subpattern itself had a zero repeat,
3234 that was OK.
3236 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3237 buffer into which it was copied (for escape processing) was still set at
3238 1024, so long lines caused crashes.
3240 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3241 "internal error: code overflow...". This applied to any character class
3242 that was followed by a possessive quantifier.
3244 15. Modified the Makefile to add libpcre.la as a prerequisite for
3245 libpcreposix.la because I was told this is needed for a parallel build to
3246 work.
3248 16. If a pattern that contained .* following optional items at the start was
3249 studied, the wrong optimizing data was generated, leading to matching
3250 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3251 matching string must start with a or b or c. The correct conclusion for
3252 this pattern is that a match can start with any character.
3255 Version 4.4 13-Aug-03
3256 ---------------------
3258 1. In UTF-8 mode, a character class containing characters with values between
3259 127 and 255 was not handled correctly if the compiled pattern was studied.
3260 In fixing this, I have also improved the studying algorithm for such
3261 classes (slightly).
3263 2. Three internal functions had redundant arguments passed to them. Removal
3264 might give a very teeny performance improvement.
3266 3. Documentation bug: the value of the capture_top field in a callout is *one
3267 more than* the number of the hightest numbered captured substring.
3269 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3270 in incorrectly linking with a previously installed version. They now link
3271 explicitly with libpcre.la.
3273 5. configure.in no longer needs to recognize Cygwin specially.
3275 6. A problem in pcre.in for Windows platforms is fixed.
3277 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3278 pcretest, it used to include the size of the study block as part of its
3279 output. Unfortunately, the structure contains a field that has a different
3280 size on different hardware architectures. This meant that the tests that
3281 showed this size failed. As the block is currently always of a fixed size,
3282 this information isn't actually particularly useful in pcretest output, so
3283 I have just removed it.
3285 8. Three pre-processor statements accidentally did not start in column 1.
3286 Sadly, there are *still* compilers around that complain, even though
3287 standard C has not required this for well over a decade. Sigh.
3289 9. In pcretest, the code for checking callouts passed small integers in the
3290 callout_data field, which is a void * field. However, some picky compilers
3291 complained about the casts involved for this on 64-bit systems. Now
3292 pcretest passes the address of the small integer instead, which should get
3293 rid of the warnings.
3295 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3296 both compile and run time, and gives an error if an invalid UTF-8 sequence
3297 is found. There is a option for disabling this check in cases where the
3298 string is known to be correct and/or the maximum performance is wanted.
3300 11. In response to a bug report, I changed one line in Makefile.in from
3302 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3303 to
3304 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3306 to look similar to other lines, but I have no way of telling whether this
3307 is the right thing to do, as I do not use Windows. No doubt I'll get told
3308 if it's wrong...
3311 Version 4.3 21-May-03
3312 ---------------------
3314 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3315 Makefile.
3317 2. Some refactoring to improve the quality of the code:
3319 (i) The utf8_table... variables are now declared "const".
3321 (ii) The code for \cx, which used the "case flipping" table to upper case
3322 lower case letters, now just substracts 32. This is ASCII-specific,
3323 but the whole concept of \cx is ASCII-specific, so it seems
3324 reasonable.
3326 (iii) PCRE was using its character types table to recognize decimal and
3327 hexadecimal digits in the pattern. This is silly, because it handles
3328 only 0-9, a-f, and A-F, but the character types table is locale-
3329 specific, which means strange things might happen. A private
3330 table is now used for this - though it costs 256 bytes, a table is
3331 much faster than multiple explicit tests. Of course, the standard
3332 character types table is still used for matching digits in subject
3333 strings against \d.
3335 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3336 ending in _t are). So I've renamed it as ESC_tee.
3338 3. The first argument for regexec() in the POSIX wrapper should have been
3339 defined as "const".
3341 4. Changed pcretest to use malloc() for its buffers so that they can be
3342 Electric Fenced for debugging.
3344 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3345 to read one or more bytes before the start of the subject string. Often this
3346 had no effect on PCRE's behaviour, but in some circumstances it could
3347 provoke a segmentation fault.
3349 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3350 to try to read one or more bytes before the start of the subject string.
3352 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3353 UTF-8 support could misbehave in various ways if the subject string
3354 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3355 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3356 back over UTF-8 characters.)
3359 Version 4.2 14-Apr-03
3360 ---------------------
3362 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3364 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3365 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3366 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3367 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3368 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3370 Note: automatic setting of the BUILD variables is not yet working
3371 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3372 compile-time but not at link-time
3373 [LINK]: use for linking executables only
3374 make different versions for Windows and non-Windows
3375 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3376 libraries
3377 [LINK_FOR_BUILD]: new variable
3378 [OBJEXT]: use throughout
3379 [EXEEXT]: use throughout
3380 <winshared>: new target
3381 <wininstall>: new target
3382 <dftables.o>: use native compiler
3383 <dftables>: use native linker
3384 <install>: handle Windows platform correctly
3385 <clean>: ditto
3386 <check>: ditto
3387 copy DLL to top builddir before testing
3389 As part of these changes, -no-undefined was removed again. This was reported
3390 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3391 in any case.
3393 3. Some tidies to get rid of compiler warnings:
3395 . In the match_data structure, match_limit was an unsigned long int, whereas
3396 match_call_count was an int. I've made them both unsigned long ints.
3398 . In pcretest the fact that a const uschar * doesn't automatically cast to
3399 a void * provoked a warning.
3401 . Turning on some more compiler warnings threw up some "shadow" variables
3402 and a few more missing casts.
3404 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3405 option, a class that contained a single character with a value between 128
3406 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3408 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3409 option, a class that contained several characters, but with at least one
3410 whose value was between 128 and 255 caused PCRE to crash.
3413 Version 4.1 12-Mar-03
3414 ---------------------
3416 1. Compiling with gcc -pedantic found a couple of places where casts were
3417 needed, and a string in dftables.c that was longer than standard compilers are
3418 required to support.
3420 2. Compiling with Sun's compiler found a few more places where the code could
3421 be tidied up in order to avoid warnings.
3423 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3424 first of these names is deprecated in the latest Autoconf in favour of the name
3425 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3426 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3427 analogy I have changed it to CFLAGS_FOR_BUILD.
3429 4. Added -no-undefined to the linking command in the Makefile, because this is
3430 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3431 linking step for the pcreposix library.
3433 5. PCRE was failing to diagnose the case of two named groups with the same
3434 name.
3436 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3437 literal character that is needed in the subject for a match, and scans along to
3438 ensure that it is present before embarking on the full matching process. This
3439 saves time in cases of nested unlimited repeats that are never going to match.
3440 Problem: the scan can take a lot of time if the subject is very long (e.g.
3441 megabytes), thus penalizing straightforward matches. It is now done only if the
3442 amount of subject to be scanned is less than 1000 bytes.
3444 7. A lesser problem with the same optimization is that it was recording the
3445 first character of an anchored pattern as "needed", thus provoking a search
3446 right along the subject, even when the first match of the pattern was going to
3447 fail. The "needed" character is now not set for anchored patterns, unless it
3448 follows something in the pattern that is of non-fixed length. Thus, it still
3449 fulfils its original purpose of finding quick non-matches in cases of nested
3450 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3453 Version 4.0 17-Feb-03
3454 ---------------------
3456 1. If a comment in an extended regex that started immediately after a meta-item
3457 extended to the end of string, PCRE compiled incorrect data. This could lead to
3458 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3460 2. Moved to autoconf 2.53 and libtool 1.4.2.
3462 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3463 the special perltest8 script is no longer needed - all the tests can be run
3464 from a single perltest script.
3466 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3467 by \s. It has now been removed in PCRE. This means it isn't recognized as
3468 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3469 class [:space:] *does* include VT, thereby creating a mess.
3471 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3472 space and tab.
3474 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3475 its new features into the main test script, reducing the number of scripts.
3477 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3478 were backward compatible, and made the (?i) apply to the whole pattern, as if
3479 /i were given. Now it behaves more logically, and applies the option setting
3480 only to what follows. PCRE has been changed to follow suit. However, if it
3481 finds options settings right at the start of the pattern, it extracts them into
3482 the global options, as before. Thus, they show up in the info data.
3484 8. Added support for the \Q...\E escape sequence. Characters in between are
3485 treated as literals. This is slightly different from Perl in that $ and @ are
3486 also handled as literals inside the quotes. In Perl, they will cause variable
3487 interpolation. Note the following examples:
3489 Pattern PCRE matches Perl matches
3491 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3492 \Qabc\$xyz\E abc\$xyz abc\$xyz
3493 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3495 For compatibility with Perl, \Q...\E sequences are recognized inside character
3496 classes as well as outside them.
3498 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3499 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3500 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3501 signed/unsigned warnings.
3503 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3504 option for pcretest, so I've replaced it by a simple function that does just
3505 that job.
3507 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3508 "pcregrep -".
3510 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3511 Java package. This provides some syntactic sugar for simple cases of what my
3512 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3513 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3514 item, you can use this simplified notation. Note that only makes sense with
3515 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3516 greediness, whatever the setting of the PCRE_UNGREEDY option.
3518 13. A change of greediness default within a pattern was not taking effect at
3519 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3520 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3521 was abstracted outside.
3523 14. PCRE now supports the \G assertion. It is true when the current matching
3524 position is at the start point of the match. This differs from \A when the
3525 starting offset is non-zero. Used with the /g option of pcretest (or similar
3526 code), it works in the same way as it does for Perl's /g option. If all
3527 alternatives of a regex begin with \G, the expression is anchored to the start
3528 match position, and the "anchored" flag is set in the compiled expression.
3530 15. Some bugs concerning the handling of certain option changes within patterns
3531 have been fixed. These applied to options other than (?ims). For example,
3532 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3533 been the other way round. Some of this was related to change 7 above.
3535 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3536 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3537 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3538 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3540 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3541 mode. Unlike ".", it always matches newline, whatever the setting of
3542 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3543 assertions. Perl allows it, but it doesn't (in general) work because it can't
3544 calculate the length of the lookbehind. At least, that's the case for Perl
3545 5.8.0 - I've been told they are going to document that it doesn't work in
3546 future.
3548 18. Added an error diagnosis for escapes that PCRE does not support: these are
3549 \L, \l, \N, \P, \p, \U, \u, and \X.
3551 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3552 reading past the end of the pattern in cases such as /[abcd/.
3554 20. PCRE was getting more memory than necessary for patterns with classes that
3555 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3557 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3558 compiling PCRE for use with Virtual Pascal.
3560 22. Small fix to the Makefile to make it work properly if the build is done
3561 outside the source tree.
3563 23. Added a new extension: a condition to go with recursion. If a conditional
3564 subpattern starts with (?(R) the "true" branch is used if recursion has
3565 happened, whereas the "false" branch is used only at the top level.
3567 24. When there was a very long string of literal characters (over 255 bytes
3568 without UTF support, over 250 bytes with UTF support), the computation of how
3569 much memory was required could be incorrect, leading to segfaults or other
3570 strange effects.
3572 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3573 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3574 there was a subsequent back reference to those brackets. This meant that, for
3575 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3576 possible to check for precisely this case. All we can do is abandon the
3577 optimization if .* occurs inside capturing brackets when there are any back
3578 references whatsoever. (See below for a better fix that came later.)
3580 26. The handling of the optimization for finding the first character of a
3581 non-anchored pattern, and for finding a character that is required later in the
3582 match were failing in some cases. This didn't break the matching; it just
3583 failed to optimize when it could. The way this is done has been re-implemented.
3585 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3587 28. Added a new feature that provides some of the functionality that Perl
3588 provides with (?{...}). The facility is termed a "callout". The way it is done
3589 in PCRE is for the caller to provide an optional function, by setting
3590 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3591 global variable. By default it is unset, which disables all calling out. To get
3592 the function called, the regex must include (?C) at appropriate points. This
3593 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3594 This provides a means of identifying different callout points. When PCRE
3595 reaches such a point in the regex, if pcre_callout has been set, the external
3596 function is called. It is provided with data in a structure called
3597 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3598 matching continues; if it returns a non-zero value, the match at the current
3599 point fails. However, backtracking will occur if possible. [This was changed
3600 later and other features added - see item 49 below.]
3602 29. pcretest is upgraded to test the callout functionality. It provides a
3603 callout function that displays information. By default, it shows the start of
3604 the match and the current position in the text. There are some new data escapes
3605 to vary what happens:
3607 \C+ in addition, show current contents of captured substrings
3608 \C- do not supply a callout function
3609 \C!n return 1 when callout number n is reached
3610 \C!n!m return 1 when callout number n is reached for the mth time
3612 30. If pcregrep was called with the -l option and just a single file name, it
3613 output "<stdin>" if a match was found, instead of the file name.
3615 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3616 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3617 pcre_exec(). This saves a malloc/free per call. The default value of
3618 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3619 when configuring.
3621 32. The default maximum size of a compiled pattern is 64K. There have been a
3622 few cases of people hitting this limit. The code now uses macros to handle the
3623 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3624 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3625 configuring. Tests 2 and 5 work only with 2-byte links because they output
3626 debugging information about compiled patterns.
3628 33. Internal code re-arrangements:
3630 (a) Moved the debugging function for printing out a compiled regex into
3631 its own source file (printint.c) and used #include to pull it into
3632 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3633 separate copies.
3635 (b) Defined the list of op-code names for debugging as a macro in
3636 internal.h so that it is next to the definition of the opcodes.
3638 (c) Defined a table of op-code lengths for simpler skipping along compiled
3639 code. This is again a macro in internal.h so that it is next to the
3640 definition of the opcodes.
3642 34. Added support for recursive calls to individual subpatterns, along the
3643 lines of Robin Houston's patch (but implemented somewhat differently).
3645 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3646 allow it to read and process whole directories in Win32. This code was
3647 contributed by Lionel Fourquaux; it has not been tested by me.
3649 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3650 used to name a group. Names consist of alphanumerics and underscores, and must
3651 be unique. Back references use the syntax (?P=name) and recursive calls use
3652 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3653 numbers. The function pcre_fullinfo() can be used after compilation to extract
3654 a name/number map. There are three relevant calls:
3656 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3657 PCRE_INFO_NAMECOUNT yields the number of entries
3658 PCRE_INFO_NAMETABLE yields a pointer to the map.
3660 The map is a vector of fixed-size entries. The size of each entry depends on
3661 the length of the longest name used. The first two bytes of each entry are the
3662 group number, most significant byte first. There follows the corresponding
3663 name, zero terminated. The names are in alphabetical order.
3665 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3666 case instead of 255. Making it the same both with and without UTF-8 support
3667 means that the same test output works with both.
3669 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3670 calling malloc() with a zero argument.
3672 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3673 optimization. I've improved things by keeping a bitmap of backreferences with
3674 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3675 fact referenced, the optimization can be applied. It is unlikely that a
3676 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3677 the match to follow \n) will appear inside brackets with a number greater than
3678 31, but if it does, any back reference > 31 suppresses the optimization.
3680 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3681 of disabling numbered capturing parentheses. Any opening parenthesis that is
3682 not followed by ? behaves as if it were followed by ?: but named parentheses
3683 can still be used for capturing (and they will acquire numbers in the usual
3684 way).
3686 41. Redesigned the return codes from the match() function into yes/no/error so
3687 that errors can be passed back from deep inside the nested calls. A malloc
3688 failure while inside a recursive subpattern call now causes the
3689 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3691 42. It is now possible to set a limit on the number of times the match()
3692 function is called in a call to pcre_exec(). This facility makes it possible to
3693 limit the amount of recursion and backtracking, though not in a directly
3694 obvious way, because the match() function is used in a number of different
3695 circumstances. The count starts from zero for each position in the subject
3696 string (for non-anchored patterns). The default limit is, for compatibility, a
3697 large number, namely 10 000 000. You can change this in two ways:
3699 (a) When configuring PCRE before making, you can use --with-match-limit=n
3700 to set a default value for the compiled library.
3702 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3703 a different value is set. See 45 below.
3705 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3707 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3708 of things that can be changed at compile time. The first argument specifies
3709 what is wanted and the second points to where the information is to be placed.
3710 The current list of available information is:
3714 The output is an integer that is set to one if UTF-8 support is available;
3715 otherwise it is set to zero.
3719 The output is an integer that it set to the value of the code that is used for
3720 newline. It is either LF (10) or CR (13).
3724 The output is an integer that contains the number of bytes used for internal
3725 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3729 The output is an integer that contains the threshold above which the POSIX
3730 interface uses malloc() for output vectors. See item 31 above.
3734 The output is an unsigned integer that contains the default limit of the number
3735 of match() calls in a pcre_exec() execution. See 42 above.
3737 44. pcretest has been upgraded by the addition of the -C option. This causes it
3738 to extract all the available output from the new pcre_config() function, and to
3739 output it. The program then exits immediately.
3741 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3742 order to support additional features. One way would have been to define
3743 pcre_exec2() (for example) with extra arguments, but this would not have been
3744 extensible, and would also have required all calls to the original function to
3745 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3746 is used for passing in "extra" data from pcre_study().
3748 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3749 contains the following fields:
3751 flags a bitmap indicating which of the following fields are set
3752 study_data opaque data from pcre_study()
3753 match_limit a way of specifying a limit on match() calls for a specific
3754 call to pcre_exec()
3755 callout_data data for callouts (see 49 below)
3757 The flag bits are also defined in pcre.h, and are
3763 The pcre_study() function now returns one of these new pcre_extra blocks, with
3764 the actual study data pointed to by the study_data field, and the
3765 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3766 before. That is, this change is entirely upwards-compatible and requires no
3767 change to existing code.
3769 If you want to pass in additional data to pcre_exec(), you can either place it
3770 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3771 block.
3773 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3774 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3775 times with different match limits, until it finds the minimum value needed for
3776 pcre_exec() to complete. The value is then output. This can be instructive; for
3777 most simple matches the number is quite small, but for pathological cases it
3778 gets very large very quickly.
3780 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3781 returns the size of the data block pointed to by the study_data field in a
3782 pcre_extra block, that is, the value that was passed as the argument to
3783 pcre_malloc() when PCRE was getting memory in which to place the information
3784 created by pcre_study(). The fourth argument should point to a size_t variable.
3785 pcretest has been extended so that this information is shown after a successful
3786 pcre_study() call when information about the compiled regex is being displayed.
3788 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3789 because what follows is always an absolute path. (Later: it turns out that this
3790 is more than cosmetic for MinGW, because it doesn't like empty path
3791 components.)
3793 49. Some changes have been made to the callout feature (see 28 above):
3795 (i) A callout function now has three choices for what it returns:
3797 0 => success, carry on matching
3798 > 0 => failure at this point, but backtrack if possible
3799 < 0 => serious error, return this value from pcre_exec()
3801 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3802 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3803 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3804 use by callout functions. It will never be used by PCRE itself.
3806 (ii) The pcre_extra structure (see 45 above) has a void * field called
3807 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3808 pcre_callout_block structure has a field of the same name. The contents of
3809 the field passed in the pcre_extra structure are passed to the callout
3810 function in the corresponding field in the callout block. This makes it
3811 easier to use the same callout-containing regex from multiple threads. For
3812 testing, the pcretest program has a new data escape
3814 \C*n pass the number n (may be negative) as callout_data
3816 If the callout function in pcretest receives a non-zero value as
3817 callout_data, it returns that value.
3819 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3820 there were some redundant $(CFLAGS) in commands that are now specified as
3821 $(LINK), which already includes $(CFLAGS).
3823 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3824 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3825 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3826 one-byte characters throughout. Note that case-insensitive matching applies
3827 only to characters whose values are less than 256. PCRE doesn't support the
3828 notion of cases for higher-valued characters.
3830 (i) A character class whose characters are all within 0-255 is handled as
3831 a bit map, and the map is inverted for negative classes. Previously, a
3832 character > 255 always failed to match such a class; however it should
3833 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3835 (ii) A negated character class with a single character < 255 is coded as
3836 "not this character" (OP_NOT). This wasn't working properly when the test
3837 character was multibyte, either singly or repeated.
3839 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3840 mode, for example: \x{100}{2,3}.
3842 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3843 singly or repeated) now correctly test multibyte characters. However,
3844 PCRE doesn't recognize any characters with values greater than 255 as
3845 digits, spaces, or word characters. Such characters always match \D, \S,
3846 and \W, and never match \d, \s, or \w.
3848 (v) Classes may now contain characters and character ranges with values
3849 greater than 255. For example: [ab\x{100}-\x{400}].
3851 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3852 PCRE in UTF-8 mode.
3854 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3855 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3856 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3857 value.)
3859 53. The single man page has become too large. I have therefore split it up into
3860 a number of separate man pages. These also give rise to individual HTML pages;
3861 these are now put in a separate directory, and there is an index.html page that
3862 lists them all. Some hyperlinking between the pages has been installed.
3864 54. Added convenience functions for handling named capturing parentheses.
3866 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3867 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3868 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3869 are faulted.
3871 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3872 calling configure. These values are used when compiling the dftables.c program
3873 which is run to generate the source of the default character tables. They
3874 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3875 you will need to set these values.
3877 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3880 Version 3.9 02-Jan-02
3881 ---------------------
3883 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3885 2. If --disable-static was given, the building process failed when trying to
3886 build pcretest and pcregrep. (For some reason it was using libtool to compile
3887 them, which is not right, as they aren't part of the library.)
3890 Version 3.8 18-Dec-01
3891 ---------------------
3893 1. The experimental UTF-8 code was completely screwed up. It was packing the
3894 bytes in the wrong order. How dumb can you get?
3897 Version 3.7 29-Oct-01
3898 ---------------------
3900 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3901 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3902 this didn't happen under Solaris 8, where I normally test things.
3904 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3905 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3906 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3907 this fix an hour or so after the initial 3.7 release.)
3910 Version 3.6 23-Oct-01
3911 ---------------------
3913 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3914 offsets passed as NULL with zero offset count.
3916 2. The config.guess and config.sub files had not been updated when I moved to
3917 the latest autoconf.
3920 Version 3.5 15-Aug-01
3921 ---------------------
3923 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3924 had been forgotten.
3926 2. By using declared but undefined structures, we can avoid using "void"
3927 definitions in pcre.h while keeping the internal definitions of the structures
3928 private.
3930 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3931 user point of view, this means that both static and shared libraries are built
3932 by default, but this can be individually controlled. More of the work of
3933 handling this static/shared cases is now inside libtool instead of PCRE's make
3934 file.
3936 4. The pcretest utility is now installed along with pcregrep because it is
3937 useful for users (to test regexs) and by doing this, it automatically gets
3938 relinked by libtool. The documentation has been turned into a man page, so
3939 there are now .1, .txt, and .html versions in /doc.
3941 5. Upgrades to pcregrep:
3942 (i) Added long-form option names like gnu grep.
3943 (ii) Added --help to list all options with an explanatory phrase.
3944 (iii) Added -r, --recursive to recurse into sub-directories.
3945 (iv) Added -f, --file to read patterns from a file.
3947 6. pcre_exec() was referring to its "code" argument before testing that
3948 argument for NULL (and giving an error if it was NULL).
3950 7. Upgraded Makefile.in to allow for compiling in a different directory from
3951 the source directory.
3953 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3954 options bits, the pointer it was passed was to an int instead of to an unsigned
3955 long int. This mattered only on 64-bit systems.
3957 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3958 generated) instead of pcre.in, which it its source. Also made the same change
3959 in several of the .c files.
3961 10. A new release of gcc defines printf() as a macro, which broke pcretest
3962 because it had an ifdef in the middle of a string argument for printf(). Fixed
3963 by using separate calls to printf().
3965 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3966 script, to force use of CR or LF instead of \n in the source. On non-Unix
3967 systems, the value can be set in config.h.
3969 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3970 absolute limit. Changed the text of the error message to make this clear, and
3971 likewise updated the man page.
3973 13. The limit of 99 on the number of capturing subpatterns has been removed.
3974 The new limit is 65535, which I hope will not be a "real" limit.
3977 Version 3.4 22-Aug-00
3978 ---------------------
3980 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3982 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3985 Version 3.3 01-Aug-00
3986 ---------------------
3988 1. If an octal character was given, but the value was greater than \377, it
3989 was not getting masked to the least significant bits, as documented. This could
3990 lead to crashes in some systems.
3992 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3993 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3995 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3996 These just pass their arguments on to (pcre_free)(), but they are provided
3997 because some uses of PCRE bind it to non-C systems that can call its functions,
3998 but cannot call free() or pcre_free() directly.
4000 4. Add "make test" as a synonym for "make check". Corrected some comments in
4001 the Makefile.
4003 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4004 Makefile.
4006 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4007 command called pgrep for grepping around the active processes.
4009 7. Added the beginnings of support for UTF-8 character strings.
4011 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4012 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4013 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4014 out for the ar command.)
4017 Version 3.2 12-May-00
4018 ---------------------
4020 This is purely a bug fixing release.
4022 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4023 of ZA. This was just one example of several cases that could provoke this bug,
4024 which was introduced by change 9 of version 2.00. The code for breaking
4025 infinite loops after an iteration that matches an empty string was't working
4026 correctly.
4028 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4029 when matched against abbab (for example). After matching an empty string, it
4030 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4031 caused it to match further down the string than it should.
4033 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4034 was there because it doesn't seem to be needed, and it causes trouble on some
4035 systems, as it is not a Standard C header. It has been removed.
4037 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4038 were reported on the Macintosh. The changes were from
4040 while ((c = *(++ptr)) != 0 && c != '\n');
4041 to
4042 while ((c = *(++ptr)) != 0 && c != '\n') ;
4044 Totally extraordinary, but if that's what it takes...
4046 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4047 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4048 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4049 assumes the way PCRE uses memmove() (always moving upwards).
4051 6. PCRE is being used in one environment where strchr() is not available. There
4052 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4053 faster code anyway.
4056 Version 3.1 09-Feb-00
4057 ---------------------
4059 The only change in this release is the fixing of some bugs in Makefile.in for
4060 the "install" target:
4062 (1) It was failing to install pcreposix.h.
4064 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4067 Version 3.0 01-Feb-00
4068 ---------------------
4070 1. Add support for the /+ modifier to perltest (to output $` like it does in
4071 pcretest).
4073 2. Add support for the /g modifier to perltest.
4075 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4076 matches null strings.
4078 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4079 pattern. Perl treats empty patterns specially - it reuses the most recent
4080 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4081 effect.
4083 5. The POSIX interface was broken in that it was just handing over the POSIX
4084 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4085 required a bigger vector, with some working space on the end. This means that
4086 the POSIX wrapper now has to get and free some memory, and copy the results.
4088 6. Added some simple autoconf support, placing the test data and the
4089 documentation in separate directories, re-organizing some of the
4090 information files, and making it build pcre-config (a GNU standard). Also added
4091 libtool support for building PCRE as a shared library, which is now the
4092 default.
4094 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4095 09 are not valid octal constants. Single digits will be used for minor values
4096 less than 10.
4098 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4099 existing programs that set these in the POSIX interface can use PCRE without
4100 modification.
4102 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4103 return all that pcre_info() returns, plus additional data. The pcre_info()
4104 function is retained for compatibility, but is considered to be obsolete.
4106 10. Added experimental recursion feature (?R) to handle one common case that
4107 Perl 5.6 will be able to do with (?p{...}).
4109 11. Added support for POSIX character classes like [:alpha:], which Perl is
4110 adopting.
4113 Version 2.08 31-Aug-99
4114 ----------------------
4116 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4117 trying to match at the startoffset position, but instead was moving forward to
4118 the next newline as if a previous match had failed.
4120 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4121 and could get into a loop if a null string was matched other than at the start
4122 of the subject.
4124 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4125 be distinguished at compile time, and for completeness also added PCRE_DATE.
4127 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4128 in GnuWin32 environments.
4131 Version 2.07 29-Jul-99
4132 ----------------------
4134 1. The documentation is now supplied in plain text form and HTML as well as in
4135 the form of man page sources.
4137 2. C++ compilers don't like assigning (void *) values to other pointer types.
4138 In particular this affects malloc(). Although there is no problem in Standard
4139 C, I've put in casts to keep C++ compilers happy.
4141 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4142 should be (const char *).
4144 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4145 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4146 However, I haven't made this a standard facility. The documentation doesn't
4147 mention it, and the Makefile doesn't support it.
4149 5. The Makefile now contains an "install" target, with editable destinations at
4150 the top of the file. The pcretest program is not installed.
4152 6. pgrep -V now gives the PCRE version number and date.
4154 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4155 causing the entire string to be ignored, instead of just the last character.
4157 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4158 non-matching string, it can take a very, very long time, even for strings of
4159 quite modest length, because of the nested recursion. PCRE now does better in
4160 some of these cases. It does this by remembering the last required literal
4161 character in the pattern, and pre-searching the subject to ensure it is present
4162 before running the real match. In other words, it applies a heuristic to detect
4163 some types of certain failure quickly, and in the above example, if presented
4164 with a string that has no trailing " it gives "no match" very quickly.
4166 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4167 other alternatives are tried instead.
4170 Version 2.06 09-Jun-99
4171 ----------------------
4173 1. Change pcretest's output for amount of store used to show just the code
4174 space, because the remainder (the data block) varies in size between 32-bit and
4175 64-bit systems.
4177 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4178 start matching at. This allows lookbehinds to work when searching for multiple
4179 occurrences in a string.
4181 3. Added additional options to pcretest for testing multiple occurrences:
4183 /+ outputs the rest of the string that follows a match
4184 /g loops for multiple occurrences, using the new startoffset argument
4185 /G loops for multiple occurrences by passing an incremented pointer
4187 4. PCRE wasn't doing the "first character" optimization for patterns starting
4188 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4189 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4190 the letter 'x'. On long subject strings, this gives a significant speed-up.
4193 Version 2.05 21-Apr-99
4194 ----------------------
4196 1. Changed the type of magic_number from int to long int so that it works
4197 properly on 16-bit systems.
4199 2. Fixed a bug which caused patterns starting with .* not to work correctly
4200 when the subject string contained newline characters. PCRE was assuming
4201 anchoring for such patterns in all cases, which is not correct because .* will
4202 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4203 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4204 must be retried after every newline in the subject.
4207 Version 2.04 18-Feb-99
4208 ----------------------
4210 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4211 computation of the store needed to hold the pattern was incorrect (too large).
4212 If such patterns were nested a few deep, this could multiply and become a real
4213 problem.
4215 2. Added /M option to pcretest to show the memory requirement of a specific
4216 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4218 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4219 compiled in such a way that the backtracking after subsequent failure was
4220 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4221 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4224 Version 2.03 02-Feb-99
4225 ----------------------
4227 1. Fixed typo and small mistake in man page.
4229 2. Added 4th condition (GPL supersedes if conflict) and created separate
4230 LICENCE file containing the conditions.
4232 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4233 Perl, that is the internal \ allows the delimiter to be included in the
4234 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4235 the final delimiter, add \ to the end of the pattern (to test the error).
4237 4. Added the convenience functions for extracting substrings after a successful
4238 match. Updated pcretest to make it able to test these functions.
4241 Version 2.02 14-Jan-99
4242 ----------------------
4244 1. Initialized the working variables associated with each extraction so that
4245 their saving and restoring doesn't refer to uninitialized store.
4247 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4248 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4249 fix the problem.
4251 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4252 calls, and wasn't printing the correct value for compiling calls. Increased the
4253 default value of LOOPREPEAT, and the number of significant figures in the
4254 times.
4256 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4258 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4259 a building problem on Windows NT with a FAT file system.
4262 Version 2.01 21-Oct-98
4263 ----------------------
4265 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4266 to character tables built by pcre_maketables() in the current locale. If NULL
4267 is passed, the default tables are used.
4270 Version 2.00 24-Sep-98
4271 ----------------------
4273 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4274 it any more.
4276 2. Allow quantification of (?>) groups, and make it work correctly.
4278 3. The first character computation wasn't working for (?>) groups.
4280 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4281 end of the subject) and add 5.005's \z, which really does match only at the
4282 very end of the subject.
4284 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4286 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4287 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4288 localized options. All options to pcre_study() were also removed.
4290 7. Add other new features from 5.005:
4292 $(?<= positive lookbehind
4293 $(?<! negative lookbehind
4294 (?imsx-imsx) added the unsetting capability
4295 such a setting is global if at outer level; local otherwise
4296 (?imsx-imsx:) non-capturing groups with option setting
4297 (?(cond)re|re) conditional pattern matching
4299 A backreference to itself in a repeated group matches the previous
4300 captured string.
4302 8. General tidying up of studying (both automatic and via "study")
4303 consequential on the addition of new assertions.
4305 9. As in 5.005, unlimited repeated groups that could match an empty substring
4306 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4307 runtime if any iteration does actually match an empty substring.
4309 10. Include the RunTest script in the distribution.
4311 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4312 discrepancies, some of which were old and were also with respect to 5.004. They
4313 have now been fixed.
4316 Version 1.09 28-Apr-98
4317 ----------------------
4319 1. A negated single character class followed by a quantifier with a minimum
4320 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4321 program crashes, or just wrong answers. This did not apply to negated classes
4322 containing more than one character, or to minima other than one.
4325 Version 1.08 27-Mar-98
4326 ----------------------
4328 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4330 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4331 latter must appear before anything that relies on it in the pattern.
4334 Version 1.07 16-Feb-98
4335 ----------------------
4337 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4338 repeat of a potentially empty string).
4341 Version 1.06 23-Jan-98
4342 ----------------------
4344 1. Added Markus Oberhumer's little patches for C++.
4346 2. Literal strings longer than 255 characters were broken.
4349 Version 1.05 23-Dec-97
4350 ----------------------
4352 1. Negated character classes containing more than one character were failing if
4353 PCRE_CASELESS was set at run time.
4356 Version 1.04 19-Dec-97
4357 ----------------------
4359 1. Corrected the man page, where some "const" qualifiers had been omitted.
4361 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4362 input syntax.
4364 3. Fixed memory leak which occurred when a regex with back references was
4365 matched with an offsets vector that wasn't big enough. The temporary memory
4366 that is used in this case wasn't being freed if the match failed.
4368 4. Tidied pcretest to ensure it frees memory that it gets.
4370 5. Temporary memory was being obtained in the case where the passed offsets
4371 vector was exactly big enough.
4373 6. Corrected definition of offsetof() from change 5 below.
4375 7. I had screwed up change 6 below and broken the rules for the use of
4376 setjmp(). Now fixed.
4379 Version 1.03 18-Dec-97
4380 ----------------------
4382 1. A erroneous regex with a missing opening parenthesis was correctly
4383 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4384 on some systems.
4386 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4387 it was reported that one broken compiler failed on the former because "code" is
4388 also an independent variable.
4390 3. The erroneous regex a[]b caused an array overrun reference.
4392 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4393 fail on data ending with that character. (It was going on too far, and checking
4394 the next character, typically a binary zero.) This was specific to the
4395 optimized code for single-character negative classes.
4397 5. Added a contributed patch from the TIN world which does the following:
4399 + Add an undef for memmove, in case the the system defines a macro for it.
4401 + Add a definition of offsetof(), in case there isn't one. (I don't know
4402 the reason behind this - offsetof() is part of the ANSI standard - but
4403 it does no harm).
4405 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4406 most of the places where whitespace preceded '#'. I have given up and
4407 allowed the remaining 2 cases to be at the margin.
4409 + Rename some variables in pcre to eliminate shadowing. This seems very
4410 pedantic, but does no harm, of course.
4412 6. Moved the call to setjmp() into its own function, to get rid of warnings
4413 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4415 7. Constructs such as \d{8,} were compiling into the equivalent of
4416 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4417 outcome, but in this particular case used more store than had been allocated,
4418 which caused the bug to be discovered because it threw up an internal error.
4420 8. The debugging code in both pcre and pcretest for outputting the compiled
4421 form of a regex was going wrong in the case of back references followed by
4422 curly-bracketed repeats.
4425 Version 1.02 12-Dec-97
4426 ----------------------
4428 1. Typos in pcre.3 and comments in the source fixed.
4430 2. Applied a contributed patch to get rid of places where it used to remove
4431 'const' from variables, and fixed some signed/unsigned and uninitialized
4432 variable warnings.
4434 3. Added the "runtest" target to Makefile.
4436 4. Set default compiler flag to -O2 rather than just -O.
4439 Version 1.01 19-Nov-97
4440 ----------------------
4442 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4443 like /([ab]*)*/, that is, for classes with more than one character in them.
4445 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4446 as /((?>a*))*/ (a PCRE_EXTRA facility).
4449 Version 1.00 18-Nov-97
4450 ----------------------
4452 1. Added compile-time macros to support systems such as SunOS4 which don't have
4453 memmove() or strerror() but have other things that can be used instead.
4455 2. Arranged that "make clean" removes the executables.
4458 Version 0.99 27-Oct-97
4459 ----------------------
4461 1. Fixed bug in code for optimizing classes with only one character. It was
4462 initializing a 32-byte map regardless, which could cause it to run off the end
4463 of the memory it had got.
4465 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4468 Version 0.98 22-Oct-97
4469 ----------------------
4471 1. Fixed bug in code for handling temporary memory usage when there are more
4472 back references than supplied space in the ovector. This could cause segfaults.
4475 Version 0.97 21-Oct-97
4476 ----------------------
4478 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4480 2. Optimized negated single characters not to use a bit map.
4482 3. Brought error texts together as macro definitions; clarified some of them;
4483 fixed one that was wrong - it said "range out of order" when it meant "invalid
4484 escape sequence".
4486 4. Changed some char * arguments to const char *.
4488 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4490 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4491 pcretest.
4494 Version 0.96 16-Oct-97
4495 ----------------------
4497 1. Added a simple "pgrep" utility to the distribution.
4499 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4500 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4501 where "ddd" means "one or more decimal digits".
4503 3. Fixed serious bug. If a pattern had a back reference, but the call to
4504 pcre_exec() didn't supply a large enough ovector to record the related
4505 identifying subpattern, the match always failed. PCRE now remembers the number
4506 of the largest back reference, and gets some temporary memory in which to save
4507 the offsets during matching if necessary, in order to ensure that
4508 backreferences always work.
4510 4. Increased the compatibility with Perl in a number of ways:
4512 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4513 to request this handling. The option can be set at compile or exec time.
4515 (b) $ matches before a terminating newline by default; an option
4516 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4517 mode). The option can be set at compile or exec time.
4519 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4520 the same as Perl's. If the decimal number it represents is less than 10
4521 or there aren't that many previous left capturing parentheses, an octal
4522 escape is read. Inside a character class, it's always an octal escape,
4523 even if it is a single digit.
4525 (d) An escaped but undefined alphabetic character is taken as a literal,
4526 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4527 escapes.
4529 (e) {0} is now permitted. (The previous item is removed from the compiled
4530 pattern).
4532 5. Changed all the names of code files so that the basic parts are no longer
4533 than 10 characters, and abolished the teeny "globals.c" file.
4535 6. Changed the handling of character classes; they are now done with a 32-byte
4536 bit map always.
4538 7. Added the -d and /D options to pcretest to make it possible to look at the
4539 internals of compilation without having to recompile pcre.
4542 Version 0.95 23-Sep-97
4543 ----------------------
4545 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4546 \x20 at the start of a run of normal characters. These were being treated as
4547 real characters, instead of the source characters being re-checked.
4550 Version 0.94 18-Sep-97
4551 ----------------------
4553 1. The functions are now thread-safe, with the caveat that the global variables
4554 containing pointers to malloc() and free() or alternative functions are the
4555 same for all threads.
4557 2. Get pcre_study() to generate a bitmap of initial characters for non-
4558 anchored patterns when this is possible, and use it if passed to pcre_exec().
4561 Version 0.93 15-Sep-97
4562 ----------------------
4564 1. /(b)|(:+)/ was computing an incorrect first character.
4566 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4567 but not actually doing anything yet.
4569 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4570 as Perl does (e.g. [-az] or [az-]).
4572 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4573 all possible positions.
4575 5. Split up into different modules to avoid including unneeded functions in a
4576 compiled binary. However, compile and exec are still in one module. The "study"
4577 function is split off.
4579 6. The character tables are now in a separate module whose source is generated
4580 by an auxiliary program - but can then be edited by hand if required. There are
4581 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4582 toupper() in the code.
4584 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4585 make them global. Abolish the function for setting them, as the caller can now
4586 set them directly.
4589 Version 0.92 11-Sep-97
4590 ----------------------
4592 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4593 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4595 2. Caseless matching was not working in character classes if the characters in
4596 the pattern were in upper case.
4598 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4600 4. Make PCRE_ANCHORED public and accept as a compile option.
4602 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4603 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4604 pass them.
4606 6. Give an error if bad option bits passed at compile or run time.
4608 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4609 pcretest to cause it to pass that flag.
4611 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4612 options, and the first character, if set.
4614 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4617 Version 0.91 10-Sep-97
4618 ----------------------
4620 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4621 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4623 2. PCRE was looping on encountering an indefinitely repeated back reference to
4624 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4625 Perl does - treats the match as successful.
4627 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5