ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 197 - (show annotations)
Tue Jul 31 10:50:18 2007 UTC (8 years, 3 months ago) by ph10
File size: 132375 byte(s)
Abolish the NULLWSLIMIT error at the expense of using more stack when an 
unlimited repeat could match an empty string. Also, doc tidies for a test 
1 ChangeLog for PCRE
2 ------------------
4 Version 7.3 05-Jul-07
5 ---------------------
7 1. In the rejigging of the build system that eventually resulted in 7.1, the
8 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
9 brackets there is not right, since it causes compilers to look for an
10 installed pcre.h, not the version that is in the source that is being
11 compiled (which of course may be different). I have changed it back to:
13 #include "pcre.h"
15 I have a vague recollection that the change was concerned with compiling in
16 different directories, but in the new build system, that is taken care of
17 by the VPATH setting the Makefile.
19 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
20 when the subject happened to end in the byte 0x85 (e.g. if the last
21 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
22 characters but of course it shouldn't be taken as a newline when it is part
23 of another character. The bug was that, for an unlimited repeat of . in
24 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
25 characters when looking for a newline.
27 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
29 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
30 in debug output.
32 5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp
33 unittest when running under MinGW.
35 6. ESC_K was left out of the EBCDIC table.
37 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
38 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
39 limit also applies to "virtual nesting" when a pattern is recursive, and in
40 this case 1000 isn't so big. I have been able to remove this limit at the
41 expense of backing off one optimization in certain circumstances. Normally,
42 when pcre_exec() would call its internal match() function recursively and
43 immediately return the result unconditionally, it uses a "tail recursion"
44 feature to save stack. However, when a subpattern that can match an empty
45 string has an unlimited repetition quantifier, it no longer makes this
46 optimization. That gives it a stack frame in which to save the data for
47 checking that an empty string has been matched. Previously this was taken
48 from the 1000-entry workspace that had been reserved. So now there is no
49 explicit limit, but more stack is used.
52 Version 7.2 19-Jun-07
53 ---------------------
55 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
56 which is apparently normally available under Windows.
58 2. Re-jig the pcregrep tests with different newline settings in an attempt
59 to make them independent of the local environment's newline setting.
61 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
63 4. Some of the "internals" tests were previously cut out when the link size
64 was not 2, because the output contained actual offsets. The recent new
65 "Z" feature of pcretest means that these can be cut out, making the tests
66 usable with all link sizes.
68 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
69 stack recursion. This gives a massive performance boost under BSD, but just
70 a small improvement under Linux. However, it saves one field in the frame
71 in all cases.
73 6. Added more features from the forthcoming Perl 5.10:
75 (a) (?-n) (where n is a string of digits) is a relative subroutine or
76 recursion call. It refers to the nth most recently opened parentheses.
78 (b) (?+n) is also a relative subroutine call; it refers to the nth next
79 to be opened parentheses.
81 (c) Conditions that refer to capturing parentheses can be specified
82 relatively, for example, (?(-2)... or (?(+3)...
84 (d) \K resets the start of the current match so that everything before
85 is not part of it.
87 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
89 (f) \g{name} is another synonym - part of Perl 5.10's unification of
90 reference syntax.
92 (g) (?| introduces a group in which the numbering of parentheses in each
93 alternative starts with the same number.
95 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
97 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
100 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
101 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
102 for detecting groups that can match an empty string.
104 9. A pattern with a very large number of alternatives (more than several
105 hundred) was running out of internal workspace during the pre-compile
106 phase, where pcre_compile() figures out how much memory will be needed. A
107 bit of new cunning has reduced the workspace needed for groups with
108 alternatives. The 1000-alternative test pattern now uses 12 bytes of
109 workspace instead of running out of the 4096 that are available.
111 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
113 11. Applied patch from Google to remove an optimization that didn't quite work.
114 The report of the bug said:
116 pcrecpp::RE("a*").FullMatch("aaa") matches, while
117 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
118 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
120 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
121 it matched the wrong number of bytes.
124 Version 7.1 24-Apr-07
125 ---------------------
127 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
128 that is more "standard", making use of automake and other Autotools. There
129 is some re-arrangement of the files and adjustment of comments consequent
130 on this.
132 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
133 for recursive directory scanning broke on some systems because the files
134 are not scanned in any specific order and on different systems the order
135 was different. A call to "sort" has been inserted into RunGrepTest for the
136 approprate test as a short-term fix. In the longer term there may be an
137 alternative.
139 3. I had an email from Eric Raymond about problems translating some of PCRE's
140 man pages to HTML (despite the fact that I distribute HTML pages, some
141 people do their own conversions for various reasons). The problems
142 concerned the use of low-level troff macros .br and .in. I have therefore
143 removed all such uses from the man pages (some were redundant, some could
144 be replaced by .nf/.fi pairs). The 132html script that I use to generate
145 HTML has been updated to handle .nf/.fi and to complain if it encounters
146 .br or .in.
148 4. Updated comments in configure.ac that get placed in config.h.in and also
149 arranged for config.h to be included in the distribution, with the name
150 config.h.generic, for the benefit of those who have to compile without
151 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
153 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
154 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
155 makevp.bat for the current PCRE, using the additional files
156 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
158 6. A Windows user reported a minor discrepancy with test 2, which turned out
159 to be caused by a trailing space on an input line that had got lost in his
160 copy. The trailing space was an accident, so I've just removed it.
162 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
163 that is needed.
165 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
166 as "const" (a) because they are and (b) because it helps the PHP
167 maintainers who have recently made a script to detect big data structures
168 in the php code that should be moved to the .rodata section. I remembered
169 to update Builducptable as well, so it won't revert if ucptable.h is ever
170 re-created.
172 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
173 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
174 order to be able to cut out the UTF-8 tables in the latter when UTF-8
175 support is not required. This saves 1.5-2K of code, which is important in
176 some applications.
178 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
179 so as not to refer to the tables, even though these functions will never be
180 called when UTF-8 support is disabled. Otherwise there are problems with a
181 shared library.
183 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
185 (a) It was defining its arguments as char * instead of void *.
187 (b) It was assuming that all moves were upwards in memory; this was true
188 a long time ago when I wrote it, but is no longer the case.
190 The emulated memove() is provided for those environments that have neither
191 memmove() nor bcopy(). I didn't think anyone used it these days, but that
192 is clearly not the case, as these two bugs were recently reported.
194 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
195 and Detrail to create the HTML documentation, the .txt form of the man
196 pages, and it removes trailing spaces from listed files. It also creates
197 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
198 case, it wraps all the #defines with #ifndefs. This script should be run
199 before "make dist".
201 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
202 with Unicode property support.
204 (a) For a maximizing quantifier, if the two different cases of the
205 character were of different lengths in their UTF-8 codings (there are
206 some cases like this - I found 11), and the matching function had to
207 back up over a mixture of the two cases, it incorrectly assumed they
208 were both the same length.
210 (b) When PCRE was configured to use the heap rather than the stack for
211 recursion during matching, it was not correctly preserving the data for
212 the other case of a UTF-8 character when checking ahead for a match
213 while processing a minimizing repeat. If the check also involved
214 matching a wide character, but failed, corruption could cause an
215 erroneous result when trying to check for a repeat of the original
216 character.
218 13. Some tidying changes to the testing mechanism:
220 (a) The RunTest script now detects the internal link size and whether there
221 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
222 values substituted by "configure". (The RunGrepTest script already did
223 this for UTF-8.) The configure.ac script no longer substitutes the
224 relevant variables.
226 (b) The debugging options /B and /D in pcretest show the compiled bytecode
227 with length and offset values. This means that the output is different
228 for different internal link sizes. Test 2 is skipped for link sizes
229 other than 2 because of this, bypassing the problem. Unfortunately,
230 there was also a test in test 3 (the locale tests) that used /B and
231 failed for link sizes other than 2. Rather than cut the whole test out,
232 I have added a new /Z option to pcretest that replaces the length and
233 offset values with spaces. This is now used to make test 3 independent
234 of link size. (Test 2 will be tidied up later.)
236 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
237 segmentation fault instead of returning the appropriate error message.
239 15. In multiline mode when the newline sequence was set to "any", the pattern
240 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
241 This doesn't seem right; it now treats the CRLF combination as the line
242 ending, and so does not match in that case. It's only a pattern such as ^$
243 that would hit this one: something like ^ABC$ would have failed after \r
244 and then tried again after \r\n.
246 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
247 in an attempt to make files that differ only in their line terminators
248 compare equal. This works on Linux.
250 17. Under certain error circumstances pcregrep might try to free random memory
251 as it exited. This is now fixed, thanks to valgrind.
253 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
254 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
255 was because its rules for how to advance for /g after matching an empty
256 string at the end of a line did not allow for this case. They now check for
257 it specially.
259 20. pcretest is supposed to handle patterns and data of any length, by
260 extending its buffers when necessary. It was getting this wrong when the
261 buffer for a data line had to be extended.
263 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
264 CRLF as a newline sequence.
266 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
267 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
268 I have nevertheless tidied it up.
270 23. Added some casts to kill warnings from HP-UX ia64 compiler.
272 24. Added a man page for pcre-config.
275 Version 7.0 19-Dec-06
276 ---------------------
278 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
279 moving to gcc 4.1.1.
281 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
282 sys/time.h, which is documented as needed for this function. It doesn't
283 seem to matter on Linux, but it showed up on some releases of OS X.
285 3. It seems that there are systems where bytes whose values are greater than
286 127 match isprint() in the "C" locale. The "C" locale should be the
287 default when a C program starts up. In most systems, only ASCII printing
288 characters match isprint(). This difference caused the output from pcretest
289 to vary, making some of the tests fail. I have changed pcretest so that:
291 (a) When it is outputting text in the compiled version of a pattern, bytes
292 other than 32-126 are always shown as hex escapes.
294 (b) When it is outputting text that is a matched part of a subject string,
295 it does the same, unless a different locale has been set for the match
296 (using the /L modifier). In this case, it uses isprint() to decide.
298 4. Fixed a major bug that caused incorrect computation of the amount of memory
299 required for a compiled pattern when options that changed within the
300 pattern affected the logic of the preliminary scan that determines the
301 length. The relevant options are -x, and -i in UTF-8 mode. The result was
302 that the computed length was too small. The symptoms of this bug were
303 either the PCRE error "internal error: code overflow" from pcre_compile(),
304 or a glibc crash with a message such as "pcretest: free(): invalid next
305 size (fast)". Examples of patterns that provoked this bug (shown in
306 pcretest format) are:
308 /(?-x: )/x
309 /(?x)(?-x: \s*#\s*)/
310 /((?i)[\x{c0}])/8
311 /(?i:[\x{c0}])/8
313 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
314 is now done differently.
316 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
317 wrapper classes; (b) implement a new function in the C++ scanner that is
318 more efficient than the old way of doing things because it avoids levels of
319 recursion in the regex matching; (c) add a paragraph to the documentation
320 for the FullMatch() function.
322 6. The escape sequence \n was being treated as whatever was defined as
323 "newline". Not only was this contrary to the documentation, which states
324 that \n is character 10 (hex 0A), but it also went horribly wrong when
325 "newline" was defined as CRLF. This has been fixed.
327 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
328 was being set to -1 for the "end of line" case (supposedly a value that no
329 character can have). Though this value is never used (the check for end of
330 line is "zero bytes in current character"), it caused compiler complaints.
331 I've changed it to 0xffffffff.
333 8. In pcre_version.c, the version string was being built by a sequence of
334 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
335 string (as it is for production releases) called a macro with an empty
336 argument. The C standard says the result of this is undefined. The gcc
337 compiler treats it as an empty string (which was what was wanted) but it is
338 reported that Visual C gives an error. The source has been hacked around to
339 avoid this problem.
341 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
342 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
343 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
344 of them did).
346 10. Originally, pcretest opened its input and output without "b"; then I was
347 told that "b" was needed in some environments, so it was added for release
348 5.0 to both the input and output. (It makes no difference on Unix-like
349 systems.) Later I was told that it is wrong for the input on Windows. I've
350 now abstracted the modes into two macros, to make it easier to fiddle with
351 them, and removed "b" from the input mode under Windows.
353 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
355 12. Added -help and --help to pcretest as an official way of being reminded
356 of the options.
358 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
359 and pcrecpp.cc because they annoy compilers at high warning levels.
361 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
363 15. Fixed an occurrence of == in configure.ac that should have been = (shell
364 scripts are not C programs :-) and which was not noticed because it works
365 on Linux.
367 16. pcretest is supposed to handle any length of pattern and data line (as one
368 line or as a continued sequence of lines) by extending its input buffer if
369 necessary. This feature was broken for very long pattern lines, leading to
370 a string of junk being passed to pcre_compile() if the pattern was longer
371 than about 50K.
373 17. I have done a major re-factoring of the way pcre_compile() computes the
374 amount of memory needed for a compiled pattern. Previously, there was code
375 that made a preliminary scan of the pattern in order to do this. That was
376 OK when PCRE was new, but as the facilities have expanded, it has become
377 harder and harder to keep it in step with the real compile phase, and there
378 have been a number of bugs (see for example, 4 above). I have now found a
379 cunning way of running the real compile function in a "fake" mode that
380 enables it to compute how much memory it would need, while actually only
381 ever using a few hundred bytes of working memory and without too many
382 tests of the mode. This should make future maintenance and development
383 easier. A side effect of this work is that the limit of 200 on the nesting
384 depth of parentheses has been removed (though this was never a serious
385 limitation, I suspect). However, there is a downside: pcre_compile() now
386 runs more slowly than before (30% or more, depending on the pattern). I
387 hope this isn't a big issue. There is no effect on runtime performance.
389 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
390 newline (only possible for the last line of a file) and it was a
391 pattern that set a locale (followed by /Lsomething), pcretest crashed.
393 19. Added additional timing features to pcretest. (1) The -tm option now times
394 matching only, not compiling. (2) Both -t and -tm can be followed, as a
395 separate command line item, by a number that specifies the number of
396 repeats to use when timing. The default is 50000; this gives better
397 precision, but takes uncomfortably long for very large patterns.
399 20. Extended pcre_study() to be more clever in cases where a branch of a
400 subpattern has no definite first character. For example, (a*|b*)[cd] would
401 previously give no result from pcre_study(). Now it recognizes that the
402 first character must be a, b, c, or d.
404 21. There was an incorrect error "recursive call could loop indefinitely" if
405 a subpattern (or the entire pattern) that was being tested for matching an
406 empty string contained only one non-empty item after a nested subpattern.
407 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
408 incorrectly, because the \d was being skipped in the check.
410 22. The pcretest program now has a new pattern option /B and a command line
411 option -b, which is equivalent to adding /B to every pattern. This causes
412 it to show the compiled bytecode, without the additional information that
413 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
414 is the same as /B/I).
416 23. A new optimization is now able automatically to treat some sequences such
417 as a*b as a*+b. More specifically, if something simple (such as a character
418 or a simple class like \d) has an unlimited quantifier, and is followed by
419 something that cannot possibly match the quantified thing, the quantifier
420 is automatically "possessified".
422 24. A recursive reference to a subpattern whose number was greater than 39
423 went wrong under certain circumstances in UTF-8 mode. This bug could also
424 have affected the operation of pcre_study().
426 25. Realized that a little bit of performance could be had by replacing
427 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
429 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
431 27. Possessive quantifiers such as a++ were previously implemented by turning
432 them into atomic groups such as ($>a+). Now they have their own opcodes,
433 which improves performance. This includes the automatically created ones
434 from 23 above.
436 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
437 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
438 the first matched character to be a colon. This applied both to named and
439 numbered groups.
441 29. The ucpinternal.h header file was missing its idempotency #ifdef.
443 30. I was sent a "project" file called libpcre.a.dev which I understand makes
444 building PCRE on Windows easier, so I have included it in the distribution.
446 31. There is now a check in pcretest against a ridiculously large number being
447 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
448 loop, the loop is abandoned.
450 32. Forward references to subpatterns in conditions such as (?(2)...) where
451 subpattern 2 is defined later cause pcre_compile() to search forwards in
452 the pattern for the relevant set of parentheses. This search went wrong
453 when there were unescaped parentheses in a character class, parentheses
454 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
456 33. "Subroutine" calls and backreferences were previously restricted to
457 referencing subpatterns earlier in the regex. This restriction has now
458 been removed.
460 34. Added a number of extra features that are going to be in Perl 5.10. On the
461 whole, these are just syntactic alternatives for features that PCRE had
462 previously implemented using the Python syntax or my own invention. The
463 other formats are all retained for compatibility.
465 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
466 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
467 also .NET compatible.
469 (b) A recursion or subroutine call to a named group can now be defined as
470 (?&name) as well as (?P>name).
472 (c) A backreference to a named group can now be defined as \k<name> or
473 \k'name' as well as (?P=name). The new forms, as well as being in Perl
474 5.10, are also .NET compatible.
476 (d) A conditional reference to a named group can now use the syntax
477 (?(<name>) or (?('name') as well as (?(name).
479 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
480 groups (named and numbered) that are never evaluated inline, but can be
481 called as "subroutines" from elsewhere. In effect, the DEFINE condition
482 is always false. There may be only one alternative in such a group.
484 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
485 as the simple (?(R). The condition is true only if the most recent
486 recursion is that of the given number or name. It does not search out
487 through the entire recursion stack.
489 (g) The escape \gN or \g{N} has been added, where N is a positive or
490 negative number, specifying an absolute or relative reference.
492 35. Tidied to get rid of some further signed/unsigned compiler warnings and
493 some "unreachable code" warnings.
495 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
496 things, this adds five new scripts.
498 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
499 There were also incompatibilities regarding the handling of \Q..\E inside
500 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
501 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
503 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
504 matches an empty string, and forcibly breaks the loop. There were bugs in
505 this code in non-simple cases. For a pattern such as ^(a()*)* matched
506 against aaaa the result was just "a" rather than "aaaa", for example. Two
507 separate and independent bugs (that affected different cases) have been
508 fixed.
510 39. Refactored the code to abolish the use of different opcodes for small
511 capturing bracket numbers. This is a tidy that I avoided doing when I
512 removed the limit on the number of capturing brackets for 3.5 back in 2001.
513 The new approach is not only tidier, it makes it possible to reduce the
514 memory needed to fix the previous bug (38).
516 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
517 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
518 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
519 mode.
521 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
522 report.
524 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
525 copy construction and assignment in the C++ wrapper.
527 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
528 couple of bugs that could have given wrong results in the "--newline=crlf"
529 case.
531 44. Added a number of casts and did some reorganization of signed/unsigned int
532 variables following suggestions from Dair Grant. Also renamed the variable
533 "this" as "item" because it is a C++ keyword.
535 45. Arranged for dftables to add
537 #include "pcre_internal.h"
539 to pcre_chartables.c because without it, gcc 4.x may remove the array
540 definition from the final binary if PCRE is built into a static library and
541 dead code stripping is activated.
543 46. For an unanchored pattern, if a match attempt fails at the start of a
544 newline sequence, and the newline setting is CRLF or ANY, and the next two
545 characters are CRLF, advance by two characters instead of one.
548 Version 6.7 04-Jul-06
549 ---------------------
551 1. In order to handle tests when input lines are enormously long, pcretest has
552 been re-factored so that it automatically extends its buffers when
553 necessary. The code is crude, but this _is_ just a test program. The
554 default size has been increased from 32K to 50K.
556 2. The code in pcre_study() was using the value of the re argument before
557 testing it for NULL. (Of course, in any sensible call of the function, it
558 won't be NULL.)
560 3. The memmove() emulation function in pcre_internal.h, which is used on
561 systems that lack both memmove() and bcopy() - that is, hardly ever -
562 was missing a "static" storage class specifier.
564 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
565 containing an extended class (one that cannot be represented by a bitmap
566 because it contains high-valued characters or Unicode property items, e.g.
567 [\pZ]). Almost always one would set UTF-8 mode when processing such a
568 pattern, but PCRE should not loop if you do not (it no longer does).
569 [Detail: two cases were found: (a) a repeated subpattern containing an
570 extended class; (b) a recursive reference to a subpattern that followed a
571 previous extended class. It wasn't skipping over the extended class
572 correctly when UTF-8 mode was not set.]
574 5. A negated single-character class was not being recognized as fixed-length
575 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
576 compile error "lookbehind assertion is not fixed length".
578 6. The RunPerlTest auxiliary script was showing an unexpected difference
579 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
580 write a Perl script that can interpret lines of an input file either as
581 byte characters or as UTF-8, which is what "perltest" was being required to
582 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
583 can't do is switch easily at run time between having the "use utf8;" pragma
584 or not. In the end, I fudged it by using the RunPerlTest script to insert
585 "use utf8;" explicitly for the UTF-8 tests.
587 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
588 the end of the subject string, contrary to the documentation and to what
589 Perl does. This was true of both matching functions. Now it matches only at
590 the start of the subject and immediately after *internal* newlines.
592 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
593 a pointer to an int instead of a pointer to an unsigned long int. This
594 caused problems on 64-bit systems.
596 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
597 instance of the 'standard' template library not being so standard".
599 10. There was no check on the number of named subpatterns nor the maximum
600 length of a subpattern name. The product of these values is used to compute
601 the size of the memory block for a compiled pattern. By supplying a very
602 long subpattern name and a large number of named subpatterns, the size
603 computation could be caused to overflow. This is now prevented by limiting
604 the length of names to 32 characters, and the number of named subpatterns
605 to 10,000.
607 11. Subpatterns that are repeated with specific counts have to be replicated in
608 the compiled pattern. The size of memory for this was computed from the
609 length of the subpattern and the repeat count. The latter is limited to
610 65535, but there was no limit on the former, meaning that integer overflow
611 could in principle occur. The compiled length of a repeated subpattern is
612 now limited to 30,000 bytes in order to prevent this.
614 12. Added the optional facility to have named substrings with the same name.
616 13. Added the ability to use a named substring as a condition, using the
617 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
618 are numbers (not recommended). Forward references are permitted.
620 14. Added forward references in named backreferences (if you see what I mean).
622 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
623 pattern could run off the end of the subject. For example, the pattern
624 "(?s)(.{1,5})"8 did this with the subject "ab".
626 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
627 PCRE_CASELESS was set when matching characters that were quantified with ?
628 or *.
630 17. A character class other than a single negated character that had a minimum
631 but no maximum quantifier - for example [ab]{6,} - was not handled
632 correctly by pce_dfa_exec(). It would match only one character.
634 18. A valid (though odd) pattern that looked like a POSIX character
635 class but used an invalid character after [ (for example [[,abc,]]) caused
636 pcre_compile() to give the error "Failed: internal error: code overflow" or
637 in some cases to crash with a glibc free() error. This could even happen if
638 the pattern terminated after [[ but there just happened to be a sequence of
639 letters, a binary zero, and a closing ] in the memory that followed.
641 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
642 over the years. Originally (before any Unicode support), just the bottom 8
643 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
644 output from "man perlunicode" includes this:
646 The regular expression compiler produces polymorphic opcodes. That
647 is, the pattern adapts to the data and automatically switches to
648 the Unicode character scheme when presented with Unicode data--or
649 instead uses a traditional byte scheme when presented with byte
650 data.
652 Sadly, a wide octal escape does not cause a switch, and in a string with
653 no other multibyte characters, these octal escapes are treated as before.
654 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
655 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
656 Unicode string.
658 I have not perpetrated such confusion in PCRE. Up till now, it took just
659 the bottom 8 bits, as in old Perl. I have now made octal escapes with
660 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
661 translate to the appropriate multibyte character.
663 29. Applied some refactoring to reduce the number of warnings from Microsoft
664 and Borland compilers. This has included removing the fudge introduced
665 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
666 a warning about an unused variable.
668 21. PCRE has not included VT (character 0x0b) in the set of whitespace
669 characters since release 4.0, because Perl (from release 5.004) does not.
670 [Or at least, is documented not to: some releases seem to be in conflict
671 with the documentation.] However, when a pattern was studied with
672 pcre_study() and all its branches started with \s, PCRE still included VT
673 as a possible starting character. Of course, this did no harm; it just
674 caused an unnecessary match attempt.
676 22. Removed a now-redundant internal flag bit that recorded the fact that case
677 dependency changed within the pattern. This was once needed for "required
678 byte" processing, but is no longer used. This recovers a now-scarce options
679 bit. Also moved the least significant internal flag bit to the most-
680 significant bit of the word, which was not previously used (hangover from
681 the days when it was an int rather than a uint) to free up another bit for
682 the future.
684 23. Added support for CRLF line endings as well as CR and LF. As well as the
685 default being selectable at build time, it can now be changed at runtime
686 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
687 specify that it is scanning data with non-default line endings.
689 24. Changed the definition of CXXLINK to make it agree with the definition of
690 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
692 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
693 recursions. This makes a big different to stack usage for some patterns.
695 26. If a subpattern containing a named recursion or subroutine reference such
696 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
697 the space required for the compiled pattern went wrong and gave too small a
698 value. Depending on the environment, this could lead to "Failed: internal
699 error: code overflow at offset 49" or "glibc detected double free or
700 corruption" errors.
702 27. Applied patches from Google (a) to support the new newline modes and (b) to
703 advance over multibyte UTF-8 characters in GlobalReplace.
705 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
706 difference for some implementation of PCRE in some Windows version.
708 29. Added some extra testing facilities to pcretest:
710 \q<number> in a data line sets the "match limit" value
711 \Q<number> in a data line sets the "match recursion limt" value
712 -S <number> sets the stack size, where <number> is in megabytes
714 The -S option isn't available for Windows.
717 Version 6.6 06-Feb-06
718 ---------------------
720 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
721 in pcreposix.h. I have copied the definition from pcre.h.
723 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
724 because pcre.h is no longer a built file.
726 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
727 not normally included in the compiled code.
730 Version 6.5 01-Feb-06
731 ---------------------
733 1. When using the partial match feature with pcre_dfa_exec(), it was not
734 anchoring the second and subsequent partial matches at the new starting
735 point. This could lead to incorrect results. For example, with the pattern
736 /1234/, partially matching against "123" and then "a4" gave a match.
738 2. Changes to pcregrep:
740 (a) All non-match returns from pcre_exec() were being treated as failures
741 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
742 error message is output. Some extra information is given for the
744 probably the only errors that are likely to be caused by users (by
745 specifying a regex that has nested indefinite repeats, for instance).
746 If there are more than 20 of these errors, pcregrep is abandoned.
748 (b) A binary zero was treated as data while matching, but terminated the
749 output line if it was written out. This has been fixed: binary zeroes
750 are now no different to any other data bytes.
752 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
753 used to set a locale for matching. The --locale=xxxx long option has
754 been added (no short equivalent) to specify a locale explicitly on the
755 pcregrep command, overriding the environment variables.
757 (d) When -B was used with -n, some line numbers in the output were one less
758 than they should have been.
760 (e) Added the -o (--only-matching) option.
762 (f) If -A or -C was used with -c (count only), some lines of context were
763 accidentally printed for the final match.
765 (g) Added the -H (--with-filename) option.
767 (h) The combination of options -rh failed to suppress file names for files
768 that were found from directory arguments.
770 (i) Added the -D (--devices) and -d (--directories) options.
772 (j) Added the -F (--fixed-strings) option.
774 (k) Allow "-" to be used as a file name for -f as well as for a data file.
776 (l) Added the --colo(u)r option.
778 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
779 is not present by default.
781 3. A nasty bug was discovered in the handling of recursive patterns, that is,
782 items such as (?R) or (?1), when the recursion could match a number of
783 alternatives. If it matched one of the alternatives, but subsequently,
784 outside the recursion, there was a failure, the code tried to back up into
785 the recursion. However, because of the way PCRE is implemented, this is not
786 possible, and the result was an incorrect result from the match.
788 In order to prevent this happening, the specification of recursion has
789 been changed so that all such subpatterns are automatically treated as
790 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
792 4. I had overlooked the fact that, in some locales, there are characters for
793 which isalpha() is true but neither isupper() nor islower() are true. In
794 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
795 and ordfeminine) are like this. This affected the treatment of \w and \W
796 when they appeared in character classes, but not when they appeared outside
797 a character class. The bit map for "word" characters is now created
798 separately from the results of isalnum() instead of just taking it from the
799 upper, lower, and digit maps. (Plus the underscore character, of course.)
801 5. The above bug also affected the handling of POSIX character classes such as
802 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
803 permanent tables. Instead, the bit maps for such a class were previously
804 created as the appropriate unions of the upper, lower, and digit bitmaps.
805 Now they are created by subtraction from the [[:word:]] class, which has
806 its own bitmap.
808 6. The [[:blank:]] character class matches horizontal, but not vertical space.
809 It is created by subtracting the vertical space characters (\x09, \x0a,
810 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
811 subtraction was done in the overall bitmap for a character class, meaning
812 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
813 be recognized. This bug has been fixed.
815 7. Patches from the folks at Google:
817 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
818 real life, but is still worth protecting against".
820 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
821 regular expressions".
823 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
824 have it.
826 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
827 "configure" and the latter not, in order to fix a problem somebody had
828 with compiling the Arg class on HP-UX.
830 (e) Improve the error-handling of the C++ wrapper a little bit.
832 (f) New tests for checking recursion limiting.
834 8. The pcre_memmove() function, which is used only if the environment does not
835 have a standard memmove() function (and is therefore rarely compiled),
836 contained two bugs: (a) use of int instead of size_t, and (b) it was not
837 returning a result (though PCRE never actually uses the result).
839 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
840 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
841 returned instead of calling malloc() with an overflowing number that would
842 most likely cause subsequent chaos.
844 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
846 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
847 with this option is matched, the nmatch and pmatch options of regexec() are
848 ignored.
850 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
851 provided in case anyone wants to the the POSIX interface with UTF-8
852 strings.
854 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
855 C++ linking (needed for some HP-UX environments).
857 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
858 (unused parameter) and in the pcre_printint() function (omitted "default"
859 switch label when the default is to do nothing).
861 15. Added some code to make it possible, when PCRE is compiled as a C++
862 library, to replace subject pointers for pcre_exec() with a smart pointer
863 class, thus making it possible to process discontinuous strings.
865 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
866 much the same function. They were added by different people who were trying
867 to make PCRE easy to compile on non-Unix systems. It has been suggested
868 that PCRE_EXPORT be abolished now that there is more automatic apparatus
869 for compiling on Windows systems. I have therefore replaced it with
870 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
871 defaults to "extern" for C or "extern C" for C++, which works fine on
872 Unix-like systems. It is now possible to override the value of PCRE_DATA_
873 SCOPE with something explicit in config.h. In addition:
875 (a) pcreposix.h still had just "extern" instead of either of these macros;
876 I have replaced it with PCRE_DATA_SCOPE.
878 (b) Functions such as _pcre_xclass(), which are internal to the library,
879 but external in the C sense, all had PCRE_EXPORT in their definitions.
880 This is apparently wrong for the Windows case, so I have removed it.
881 (It makes no difference on Unix-like systems.)
883 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
884 of recursive calls to match(). This is different to MATCH_LIMIT because
885 that limits the total number of calls to match(), not all of which increase
886 the depth of recursion. Limiting the recursion depth limits the amount of
887 stack (or heap if NO_RECURSE is set) that is used. The default can be set
888 when PCRE is compiled, and changed at run time. A patch from Google adds
889 this functionality to the C++ interface.
891 18. Changes to the handling of Unicode character properties:
893 (a) Updated the table to Unicode 4.1.0.
895 (b) Recognize characters that are not in the table as "Cn" (undefined).
897 (c) I revised the way the table is implemented to a much improved format
898 which includes recognition of ranges. It now supports the ranges that
899 are defined in UnicodeData.txt, and it also amalgamates other
900 characters into ranges. This has reduced the number of entries in the
901 table from around 16,000 to around 3,000, thus reducing its size
902 considerably. I realized I did not need to use a tree structure after
903 all - a binary chop search is just as efficient. Having reduced the
904 number of entries, I extended their size from 6 bytes to 8 bytes to
905 allow for more data.
907 (d) Added support for Unicode script names via properties such as \p{Han}.
909 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
910 matching that character.
912 20. When matching a repeated Unicode property with a minimum greater than zero,
913 (for example \pL{2,}), PCRE could look past the end of the subject if it
914 reached it while seeking the minimum number of characters. This could
915 happen only if some of the characters were more than one byte long, because
916 there is a check for at least the minimum number of bytes.
918 21. Refactored the implementation of \p and \P so as to be more general, to
919 allow for more different types of property in future. This has changed the
920 compiled form incompatibly. Anybody with saved compiled patterns that use
921 \p or \P will have to recompile them.
923 22. Added "Any" and "L&" to the supported property types.
925 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
926 but give a compile time error if the value is greater than 0xff.
928 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
929 accidentally not being installed or uninstalled.
931 25. The pcre.h file was built from pcre.h.in, but the only changes that were
932 made were to insert the current release number. This seemed silly, because
933 it made things harder for people building PCRE on systems that don't run
934 "configure". I have turned pcre.h into a distributed file, no longer built
935 by "configure", with the version identification directly included. There is
936 no longer a pcre.h.in file.
938 However, this change necessitated a change to the pcre-config script as
939 well. It is built from pcre-config.in, and one of the substitutions was the
940 release number. I have updated configure.ac so that ./configure now finds
941 the release number by grepping pcre.h.
943 26. Added the ability to run the tests under valgrind.
946 Version 6.4 05-Sep-05
947 ---------------------
949 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
950 "--" to be printed when multiple files were scanned, even when none of the
951 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
952 consider it to be a bug, and have restored the previous behaviour.
954 2. A couple of code tidies to get rid of compiler warnings.
956 3. The pcretest program used to cheat by referring to symbols in the library
957 whose names begin with _pcre_. These are internal symbols that are not
958 really supposed to be visible externally, and in some environments it is
959 possible to suppress them. The cheating is now confined to including
960 certain files from the library's source, which is a bit cleaner.
962 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
963 file's purpose clearer.
965 5. Reorganized pcre_ucp_findchar().
968 Version 6.3 15-Aug-05
969 ---------------------
971 1. The file libpcre.pc.in did not have general read permission in the tarball.
973 2. There were some problems when building without C++ support:
975 (a) If C++ support was not built, "make install" and "make test" still
976 tried to test it.
978 (b) There were problems when the value of CXX was explicitly set. Some
979 changes have been made to try to fix these, and ...
981 (c) --disable-cpp can now be used to explicitly disable C++ support.
983 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
984 backslash in a target when C++ was disabled. This confuses some
985 versions of "make", apparently. Using an intermediate variable solves
986 this. (Same for CPP_LOBJ.)
988 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
989 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
990 necessary on certain architectures.
992 4. Added a setting of -export-symbols-regex to the link command to remove
993 those symbols that are exported in the C sense, but actually are local
994 within the library, and not documented. Their names all begin with
995 "_pcre_". This is not a perfect job, because (a) we have to except some
996 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
997 available (and never for static libraries). I have made a note to try to
998 find a way round (a) in the future.
1001 Version 6.2 01-Aug-05
1002 ---------------------
1004 1. There was no test for integer overflow of quantifier values. A construction
1005 such as {1111111111111111} would give undefined results. What is worse, if
1006 a minimum quantifier for a parenthesized subpattern overflowed and became
1007 negative, the calculation of the memory size went wrong. This could have
1008 led to memory overwriting.
1010 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1012 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1013 operating environments where this matters.
1015 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1016 PCRE options from within the C++ wrapper.
1018 5. Named capturing subpatterns were not being correctly counted when a pattern
1019 was compiled. This caused two problems: (a) If there were more than 100
1020 such subpatterns, the calculation of the memory needed for the whole
1021 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1022 back references of the form \12, where the number was greater than 9, were
1023 not recognized as back references, even though there were sufficient
1024 previous subpatterns.
1026 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1027 versions of gcc, e.g. 2.95.4.
1030 Version 6.1 21-Jun-05
1031 ---------------------
1033 1. There was one reference to the variable "posix" in pcretest.c that was not
1034 surrounded by "#if !defined NOPOSIX".
1036 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1037 the cross-check on the old pcre_info() function, for the benefit of the
1038 cut-down version of PCRE that is currently imported into Exim.
1040 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1041 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1042 patterns that start (?i) but I don't think that matters. The use of (?i) is
1043 just an example; this all applies to the other options as well.
1045 4. Since libtool seems to echo the compile commands it is issuing, the output
1046 from "make" can be reduced a bit by putting "@" in front of each libtool
1047 compile command.
1049 5. Patch from the folks at Google for configure.in to be a bit more thorough
1050 in checking for a suitable C++ installation before trying to compile the
1051 C++ stuff. This should fix a reported problem when a compiler was present,
1052 but no suitable headers.
1054 6. The man pages all had just "PCRE" as their title. I have changed them to
1055 be the relevant file name. I have also arranged that these names are
1056 retained in the file doc/pcre.txt, which is a concatenation in text format
1057 of all the man pages except the little individual ones for each function.
1059 7. The NON-UNIX-USE file had not been updated for the different set of source
1060 files that come with release 6. I also added a few comments about the C++
1061 wrapper.
1064 Version 6.0 07-Jun-05
1065 ---------------------
1067 1. Some minor internal re-organization to help with my DFA experiments.
1069 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1070 didn't matter for the library itself when fully configured, but did matter
1071 when compiling without UCP support, or within Exim, where the ucp files are
1072 not imported.
1074 3. Refactoring of the library code to split up the various functions into
1075 different source modules. The addition of the new DFA matching code (see
1076 below) to a single monolithic source would have made it really too
1077 unwieldy, quite apart from causing all the code to be include in a
1078 statically linked application, when only some functions are used. This is
1079 relevant even without the DFA addition now that patterns can be compiled in
1080 one application and matched in another.
1082 The downside of splitting up is that there have to be some external
1083 functions and data tables that are used internally in different modules of
1084 the library but which are not part of the API. These have all had their
1085 names changed to start with "_pcre_" so that they are unlikely to clash
1086 with other external names.
1088 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1089 a different (DFA) algorithm. Although it is slower than the original
1090 function, it does have some advantages for certain types of matching
1091 problem.
1093 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1094 including restarting after a partial match.
1096 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1097 defined when compiling for Windows was sent to me. I have put it into the
1098 code, though I have no means of testing or verifying it.
1100 7. Added the pcre_refcount() auxiliary function.
1102 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1103 match before or at the first newline in the subject string. In pcretest,
1104 the /f option on a pattern can be used to set this.
1106 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1107 would behave wrongly. This has been present in PCRE since release 4.0.
1109 10. A number of changes to the pcregrep command:
1111 (a) Refactored how -x works; insert ^(...)$ instead of setting
1112 PCRE_ANCHORED and checking the length, in preparation for adding
1113 something similar for -w.
1115 (b) Added the -w (match as a word) option.
1117 (c) Refactored the way lines are read and buffered so as to have more
1118 than one at a time available.
1120 (d) Implemented a pcregrep test script.
1122 (e) Added the -M (multiline match) option. This allows patterns to match
1123 over several lines of the subject. The buffering ensures that at least
1124 8K, or the rest of the document (whichever is the shorter) is available
1125 for matching (and similarly the previous 8K for lookbehind assertions).
1127 (f) Changed the --help output so that it now says
1129 -w, --word-regex(p)
1131 instead of two lines, one with "regex" and the other with "regexp"
1132 because that confused at least one person since the short forms are the
1133 same. (This required a bit of code, as the output is generated
1134 automatically from a table. It wasn't just a text change.)
1136 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1137 option but starts with a hyphen. Could be a pattern or a path name
1138 starting with a hyphen, for instance.
1140 (h) "-" can be given as a file name to represent stdin.
1142 (i) When file names are being printed, "(standard input)" is used for
1143 the standard input, for compatibility with GNU grep. Previously
1144 "<stdin>" was used.
1146 (j) The option --label=xxx can be used to supply a name to be used for
1147 stdin when file names are being printed. There is no short form.
1149 (k) Re-factored the options decoding logic because we are going to add
1150 two more options that take data. Such options can now be given in four
1151 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1153 (l) Added the -A, -B, and -C options for requesting that lines of context
1154 around matches be printed.
1156 (m) Added the -L option to print the names of files that do not contain
1157 any matching lines, that is, the complement of -l.
1159 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1160 continue to scan other files.
1162 (o) The -s option was incorrectly implemented. For compatibility with other
1163 greps, it now suppresses the error message for a non-existent or non-
1164 accessible file (but not the return code). There is a new option called
1165 -q that suppresses the output of matching lines, which was what -s was
1166 previously doing.
1168 (p) Added --include and --exclude options to specify files for inclusion
1169 and exclusion when recursing.
1171 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1172 Hopefully, it now does.
1174 12. Missing cast in pcre_study().
1176 13. Added an "uninstall" target to the makefile.
1178 14. Replaced "extern" in the function prototypes in Makefile.in with
1179 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1180 world, but is set differently for Windows.
1182 15. Added a second compiling function called pcre_compile2(). The only
1183 difference is that it has an extra argument, which is a pointer to an
1184 integer error code. When there is a compile-time failure, this is set
1185 non-zero, in addition to the error test pointer being set to point to an
1186 error message. The new argument may be NULL if no error number is required
1187 (but then you may as well call pcre_compile(), which is now just a
1188 wrapper). This facility is provided because some applications need a
1189 numeric error indication, but it has also enabled me to tidy up the way
1190 compile-time errors are handled in the POSIX wrapper.
1192 16. Added VPATH=.libs to the makefile; this should help when building with one
1193 prefix path and installing with another. (Or so I'm told by someone who
1194 knows more about this stuff than I do.)
1196 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1197 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1198 match everything, including newlines. This is not POSIX-compatible, but
1199 somebody wanted the feature. From pcretest it can be activated by using
1200 both the P and the s flags.
1202 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1204 19. libpcre.pc was being incorrectly installed as executable.
1206 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1207 it now also looks for '\r' so that it will work unmodified on Windows.
1209 21. Added Google's contributed C++ wrapper to the distribution.
1211 22. Added some untidy missing memory free() calls in pcretest, to keep
1212 Electric Fence happy when testing.
1216 Version 5.0 13-Sep-04
1217 ---------------------
1219 1. Internal change: literal characters are no longer packed up into items
1220 containing multiple characters in a single byte-string. Each character
1221 is now matched using a separate opcode. However, there may be more than one
1222 byte in the character in UTF-8 mode.
1224 2. The pcre_callout_block structure has two new fields: pattern_position and
1225 next_item_length. These contain the offset in the pattern to the next match
1226 item, and its length, respectively.
1228 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1229 insertion of callouts before each pattern item. Added the /C option to
1230 pcretest to make use of this.
1232 4. On the advice of a Windows user, the lines
1234 #if defined(_WIN32) || defined(WIN32)
1235 _setmode( _fileno( stdout ), 0x8000 );
1236 #endif /* defined(_WIN32) || defined(WIN32) */
1238 have been added to the source of pcretest. This apparently does useful
1239 magic in relation to line terminators.
1241 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1242 for the benefit of those environments where the "b" makes a difference.
1244 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1245 to know about it. I have put a hack into configure.in that adds in code
1246 to set GCC=yes if CC=icc. This seems to end up at a point in the
1247 generated configure script that is early enough to affect the setting of
1248 compiler options, which is what is needed, but I have no means of testing
1249 whether it really works. (The user who reported this had patched the
1250 generated configure script, which of course I cannot do.)
1252 LATER: After change 22 below (new libtool files), the configure script
1253 seems to know about icc (and also ecc). Therefore, I have commented out
1254 this hack in configure.in.
1256 7. Added support for pkg-config (2 patches were sent in).
1258 8. Negated POSIX character classes that used a combination of internal tables
1259 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1260 [[:^ascii]]. Typically, they would match almost any characters. The other
1261 POSIX classes were not broken in this way.
1263 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1264 to find the match, as PCRE was deluded into thinking that the match had to
1265 start at the start point or following a newline. The same bug applied to
1266 patterns with negative forward assertions or any backward assertions
1267 preceding ".*" at the start, unless the pattern required a fixed first
1268 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1270 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1271 starting at the last subject character, bytes beyond the end of the subject
1272 string were read.
1274 11. Renamed the variable "class" as "classbits" to make life easier for C++
1275 users. (Previously there was a macro definition, but it apparently wasn't
1276 enough.)
1278 12. Added the new field "tables" to the extra data so that tables can be passed
1279 in at exec time, or the internal tables can be re-selected. This allows
1280 a compiled regex to be saved and re-used at a later time by a different
1281 program that might have everything at different addresses.
1283 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1284 -R library as well as a -L library.
1286 14. The debugging options of pcretest (-d on the command line or D on a
1287 pattern) showed incorrect output for anything following an extended class
1288 that contained multibyte characters and which was followed by a quantifier.
1290 15. Added optional support for general category Unicode character properties
1291 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1292 support. It adds about 90K to the size of the library. The meanings of the
1293 inbuilt class escapes such as \d and \s have NOT been changed.
1295 16. Updated pcredemo.c to include calls to free() to release the memory for the
1296 compiled pattern.
1298 17. The generated file chartables.c was being created in the source directory
1299 instead of in the building directory. This caused the build to fail if the
1300 source directory was different from the building directory, and was
1301 read-only.
1303 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1304 file. No doubt somebody will tell me if they don't make sense... Also added
1305 Dan Mooney's comments about building on OpenVMS.
1307 19. Added support for partial matching via the PCRE_PARTIAL option for
1308 pcre_exec() and the \P data escape in pcretest.
1310 20. Extended pcretest with 3 new pattern features:
1312 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1313 write the compiled pattern to the file whose name is "rest-of-line".
1314 This is a straight binary dump of the data, with the saved pointer to
1315 the character tables forced to be NULL. The study data, if any, is
1316 written too. After writing, pcretest reads a new pattern.
1318 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1319 compiled pattern from the given file. There must not be any
1320 occurrences of "<" in the file name (pretty unlikely); if there are,
1321 pcretest will instead treat the initial "<" as a pattern delimiter.
1322 After reading in the pattern, pcretest goes on to read data lines as
1323 usual.
1325 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1326 and 16-bit fields in a compiled pattern, to simulate a pattern that
1327 was compiled on a host of opposite endianness.
1329 21. The pcre-exec() function can now cope with patterns that were compiled on
1330 hosts of opposite endianness, with this restriction:
1332 As for any compiled expression that is saved and used later, the tables
1333 pointer field cannot be preserved; the extra_data field in the arguments
1334 to pcre_exec() should be used to pass in a tables address if a value
1335 other than the default internal tables were used at compile time.
1337 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1338 now diagnosed as an error. Previously, most of the time, a negative number
1339 would have been treated as zero, but if in addition "ovector" was passed as
1340 NULL, a crash could occur.
1342 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1343 new versions from the libtool 1.5 distribution (the last one is a copy of
1344 a file called libtool.m4). This seems to have fixed the need to patch
1345 "configure" to support Darwin 1.3 (which I used to do). However, I still
1346 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1347 workstation).
1349 24. Changed the PCRE licence to be the more standard "BSD" licence.
1352 Version 4.5 01-Dec-03
1353 ---------------------
1355 1. There has been some re-arrangement of the code for the match() function so
1356 that it can be compiled in a version that does not call itself recursively.
1357 Instead, it keeps those local variables that need separate instances for
1358 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1359 needs to "recurse". Keeping track of where control must go is done by means
1360 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1361 hide most of the details from the main code, and operates only if
1362 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1363 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1364 operating.
1366 To make it easier for callers to provide specially tailored get/free
1367 functions for this usage, two new functions, pcre_stack_malloc, and
1368 pcre_stack_free, are used. They are always called in strict stacking order,
1369 and the size of block requested is always the same.
1371 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1372 PCRE has been compiled to use the stack or the heap for recursion. The
1373 -C option of pcretest uses this to show which version is compiled.
1375 A new data escape \S, is added to pcretest; it causes the amounts of store
1376 obtained and freed by both kinds of malloc/free at match time to be added
1377 to the output.
1379 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1380 what's available on my current Linux desktop machine.
1382 3. When matching a UTF-8 string, the test for a valid string at the start has
1383 been extended. If start_offset is not zero, PCRE now checks that it points
1384 to a byte that is the start of a UTF-8 character. If not, it returns
1385 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1386 this is necessary because there may be backward assertions in the pattern.
1387 When matching the same subject several times, it may save resources to use
1388 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1390 4. The code for checking the validity of UTF-8 strings has been tightened so
1391 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1392 containing "overlong sequences".
1394 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1395 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1396 should have been "&", but it just so happened that all the cases this let
1397 through by mistake were picked up later in the function.
1399 6. I had used a variable called "isblank" - this is a C99 function, causing
1400 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1402 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1403 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1404 so the version is shown; (c) stop "make test" echoing "./RunTest".
1406 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1408 9. The prototype for memmove() for systems that don't have it was using
1409 size_t, but the inclusion of the header that defines size_t was later. I've
1410 moved the #includes for the C headers earlier to avoid this.
1412 10. Added some adjustments to the code to make it easier to compiler on certain
1413 special systems:
1415 (a) Some "const" qualifiers were missing.
1416 (b) Added the macro EXPORT before all exported functions; by default this
1417 is defined to be empty.
1418 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1419 that it reads its output file name as an argument instead of writing
1420 to the standard output and assuming this can be redirected.
1422 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1423 class containing characters with values greater than 255, PCRE compilation
1424 went into a loop.
1426 12. A recursive reference to a subpattern that was within another subpattern
1427 that had a minimum quantifier of zero caused PCRE to crash. For example,
1428 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1429 recursion. If the recursively-called subpattern itself had a zero repeat,
1430 that was OK.
1432 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1433 buffer into which it was copied (for escape processing) was still set at
1434 1024, so long lines caused crashes.
1436 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1437 "internal error: code overflow...". This applied to any character class
1438 that was followed by a possessive quantifier.
1440 15. Modified the Makefile to add libpcre.la as a prerequisite for
1441 libpcreposix.la because I was told this is needed for a parallel build to
1442 work.
1444 16. If a pattern that contained .* following optional items at the start was
1445 studied, the wrong optimizing data was generated, leading to matching
1446 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1447 matching string must start with a or b or c. The correct conclusion for
1448 this pattern is that a match can start with any character.
1451 Version 4.4 13-Aug-03
1452 ---------------------
1454 1. In UTF-8 mode, a character class containing characters with values between
1455 127 and 255 was not handled correctly if the compiled pattern was studied.
1456 In fixing this, I have also improved the studying algorithm for such
1457 classes (slightly).
1459 2. Three internal functions had redundant arguments passed to them. Removal
1460 might give a very teeny performance improvement.
1462 3. Documentation bug: the value of the capture_top field in a callout is *one
1463 more than* the number of the hightest numbered captured substring.
1465 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1466 in incorrectly linking with a previously installed version. They now link
1467 explicitly with libpcre.la.
1469 5. configure.in no longer needs to recognize Cygwin specially.
1471 6. A problem in pcre.in for Windows platforms is fixed.
1473 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1474 pcretest, it used to include the size of the study block as part of its
1475 output. Unfortunately, the structure contains a field that has a different
1476 size on different hardware architectures. This meant that the tests that
1477 showed this size failed. As the block is currently always of a fixed size,
1478 this information isn't actually particularly useful in pcretest output, so
1479 I have just removed it.
1481 8. Three pre-processor statements accidentally did not start in column 1.
1482 Sadly, there are *still* compilers around that complain, even though
1483 standard C has not required this for well over a decade. Sigh.
1485 9. In pcretest, the code for checking callouts passed small integers in the
1486 callout_data field, which is a void * field. However, some picky compilers
1487 complained about the casts involved for this on 64-bit systems. Now
1488 pcretest passes the address of the small integer instead, which should get
1489 rid of the warnings.
1491 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1492 both compile and run time, and gives an error if an invalid UTF-8 sequence
1493 is found. There is a option for disabling this check in cases where the
1494 string is known to be correct and/or the maximum performance is wanted.
1496 11. In response to a bug report, I changed one line in Makefile.in from
1498 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1499 to
1500 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1502 to look similar to other lines, but I have no way of telling whether this
1503 is the right thing to do, as I do not use Windows. No doubt I'll get told
1504 if it's wrong...
1507 Version 4.3 21-May-03
1508 ---------------------
1510 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1511 Makefile.
1513 2. Some refactoring to improve the quality of the code:
1515 (i) The utf8_table... variables are now declared "const".
1517 (ii) The code for \cx, which used the "case flipping" table to upper case
1518 lower case letters, now just substracts 32. This is ASCII-specific,
1519 but the whole concept of \cx is ASCII-specific, so it seems
1520 reasonable.
1522 (iii) PCRE was using its character types table to recognize decimal and
1523 hexadecimal digits in the pattern. This is silly, because it handles
1524 only 0-9, a-f, and A-F, but the character types table is locale-
1525 specific, which means strange things might happen. A private
1526 table is now used for this - though it costs 256 bytes, a table is
1527 much faster than multiple explicit tests. Of course, the standard
1528 character types table is still used for matching digits in subject
1529 strings against \d.
1531 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1532 ending in _t are). So I've renamed it as ESC_tee.
1534 3. The first argument for regexec() in the POSIX wrapper should have been
1535 defined as "const".
1537 4. Changed pcretest to use malloc() for its buffers so that they can be
1538 Electric Fenced for debugging.
1540 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1541 to read one or more bytes before the start of the subject string. Often this
1542 had no effect on PCRE's behaviour, but in some circumstances it could
1543 provoke a segmentation fault.
1545 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1546 to try to read one or more bytes before the start of the subject string.
1548 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1549 UTF-8 support could misbehave in various ways if the subject string
1550 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1551 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1552 back over UTF-8 characters.)
1555 Version 4.2 14-Apr-03
1556 ---------------------
1558 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1560 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1561 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1562 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1563 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1564 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1566 Note: automatic setting of the BUILD variables is not yet working
1567 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1568 compile-time but not at link-time
1569 [LINK]: use for linking executables only
1570 make different versions for Windows and non-Windows
1571 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1572 libraries
1573 [LINK_FOR_BUILD]: new variable
1574 [OBJEXT]: use throughout
1575 [EXEEXT]: use throughout
1576 <winshared>: new target
1577 <wininstall>: new target
1578 <dftables.o>: use native compiler
1579 <dftables>: use native linker
1580 <install>: handle Windows platform correctly
1581 <clean>: ditto
1582 <check>: ditto
1583 copy DLL to top builddir before testing
1585 As part of these changes, -no-undefined was removed again. This was reported
1586 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1587 in any case.
1589 3. Some tidies to get rid of compiler warnings:
1591 . In the match_data structure, match_limit was an unsigned long int, whereas
1592 match_call_count was an int. I've made them both unsigned long ints.
1594 . In pcretest the fact that a const uschar * doesn't automatically cast to
1595 a void * provoked a warning.
1597 . Turning on some more compiler warnings threw up some "shadow" variables
1598 and a few more missing casts.
1600 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1601 option, a class that contained a single character with a value between 128
1602 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1604 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1605 option, a class that contained several characters, but with at least one
1606 whose value was between 128 and 255 caused PCRE to crash.
1609 Version 4.1 12-Mar-03
1610 ---------------------
1612 1. Compiling with gcc -pedantic found a couple of places where casts were
1613 needed, and a string in dftables.c that was longer than standard compilers are
1614 required to support.
1616 2. Compiling with Sun's compiler found a few more places where the code could
1617 be tidied up in order to avoid warnings.
1619 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1620 first of these names is deprecated in the latest Autoconf in favour of the name
1621 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1622 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1623 analogy I have changed it to CFLAGS_FOR_BUILD.
1625 4. Added -no-undefined to the linking command in the Makefile, because this is
1626 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1627 linking step for the pcreposix library.
1629 5. PCRE was failing to diagnose the case of two named groups with the same
1630 name.
1632 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1633 literal character that is needed in the subject for a match, and scans along to
1634 ensure that it is present before embarking on the full matching process. This
1635 saves time in cases of nested unlimited repeats that are never going to match.
1636 Problem: the scan can take a lot of time if the subject is very long (e.g.
1637 megabytes), thus penalizing straightforward matches. It is now done only if the
1638 amount of subject to be scanned is less than 1000 bytes.
1640 7. A lesser problem with the same optimization is that it was recording the
1641 first character of an anchored pattern as "needed", thus provoking a search
1642 right along the subject, even when the first match of the pattern was going to
1643 fail. The "needed" character is now not set for anchored patterns, unless it
1644 follows something in the pattern that is of non-fixed length. Thus, it still
1645 fulfils its original purpose of finding quick non-matches in cases of nested
1646 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1649 Version 4.0 17-Feb-03
1650 ---------------------
1652 1. If a comment in an extended regex that started immediately after a meta-item
1653 extended to the end of string, PCRE compiled incorrect data. This could lead to
1654 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1656 2. Moved to autoconf 2.53 and libtool 1.4.2.
1658 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1659 the special perltest8 script is no longer needed - all the tests can be run
1660 from a single perltest script.
1662 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1663 by \s. It has now been removed in PCRE. This means it isn't recognized as
1664 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1665 class [:space:] *does* include VT, thereby creating a mess.
1667 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1668 space and tab.
1670 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1671 its new features into the main test script, reducing the number of scripts.
1673 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1674 were backward compatible, and made the (?i) apply to the whole pattern, as if
1675 /i were given. Now it behaves more logically, and applies the option setting
1676 only to what follows. PCRE has been changed to follow suit. However, if it
1677 finds options settings right at the start of the pattern, it extracts them into
1678 the global options, as before. Thus, they show up in the info data.
1680 8. Added support for the \Q...\E escape sequence. Characters in between are
1681 treated as literals. This is slightly different from Perl in that $ and @ are
1682 also handled as literals inside the quotes. In Perl, they will cause variable
1683 interpolation. Note the following examples:
1685 Pattern PCRE matches Perl matches
1687 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1688 \Qabc\$xyz\E abc\$xyz abc\$xyz
1689 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1691 For compatibility with Perl, \Q...\E sequences are recognized inside character
1692 classes as well as outside them.
1694 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1695 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1696 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1697 signed/unsigned warnings.
1699 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1700 option for pcretest, so I've replaced it by a simple function that does just
1701 that job.
1703 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1704 "pcregrep -".
1706 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1707 Java package. This provides some syntactic sugar for simple cases of what my
1708 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1709 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1710 item, you can use this simplified notation. Note that only makes sense with
1711 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1712 greediness, whatever the setting of the PCRE_UNGREEDY option.
1714 13. A change of greediness default within a pattern was not taking effect at
1715 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1716 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1717 was abstracted outside.
1719 14. PCRE now supports the \G assertion. It is true when the current matching
1720 position is at the start point of the match. This differs from \A when the
1721 starting offset is non-zero. Used with the /g option of pcretest (or similar
1722 code), it works in the same way as it does for Perl's /g option. If all
1723 alternatives of a regex begin with \G, the expression is anchored to the start
1724 match position, and the "anchored" flag is set in the compiled expression.
1726 15. Some bugs concerning the handling of certain option changes within patterns
1727 have been fixed. These applied to options other than (?ims). For example,
1728 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1729 been the other way round. Some of this was related to change 7 above.
1731 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1732 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1733 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1734 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1736 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1737 mode. Unlike ".", it always matches newline, whatever the setting of
1738 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1739 assertions. Perl allows it, but it doesn't (in general) work because it can't
1740 calculate the length of the lookbehind. At least, that's the case for Perl
1741 5.8.0 - I've been told they are going to document that it doesn't work in
1742 future.
1744 18. Added an error diagnosis for escapes that PCRE does not support: these are
1745 \L, \l, \N, \P, \p, \U, \u, and \X.
1747 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1748 reading past the end of the pattern in cases such as /[abcd/.
1750 20. PCRE was getting more memory than necessary for patterns with classes that
1751 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1753 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1754 compiling PCRE for use with Virtual Pascal.
1756 22. Small fix to the Makefile to make it work properly if the build is done
1757 outside the source tree.
1759 23. Added a new extension: a condition to go with recursion. If a conditional
1760 subpattern starts with (?(R) the "true" branch is used if recursion has
1761 happened, whereas the "false" branch is used only at the top level.
1763 24. When there was a very long string of literal characters (over 255 bytes
1764 without UTF support, over 250 bytes with UTF support), the computation of how
1765 much memory was required could be incorrect, leading to segfaults or other
1766 strange effects.
1768 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1769 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1770 there was a subsequent back reference to those brackets. This meant that, for
1771 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1772 possible to check for precisely this case. All we can do is abandon the
1773 optimization if .* occurs inside capturing brackets when there are any back
1774 references whatsoever. (See below for a better fix that came later.)
1776 26. The handling of the optimization for finding the first character of a
1777 non-anchored pattern, and for finding a character that is required later in the
1778 match were failing in some cases. This didn't break the matching; it just
1779 failed to optimize when it could. The way this is done has been re-implemented.
1781 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1783 28. Added a new feature that provides some of the functionality that Perl
1784 provides with (?{...}). The facility is termed a "callout". The way it is done
1785 in PCRE is for the caller to provide an optional function, by setting
1786 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1787 global variable. By default it is unset, which disables all calling out. To get
1788 the function called, the regex must include (?C) at appropriate points. This
1789 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1790 This provides a means of identifying different callout points. When PCRE
1791 reaches such a point in the regex, if pcre_callout has been set, the external
1792 function is called. It is provided with data in a structure called
1793 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1794 matching continues; if it returns a non-zero value, the match at the current
1795 point fails. However, backtracking will occur if possible. [This was changed
1796 later and other features added - see item 49 below.]
1798 29. pcretest is upgraded to test the callout functionality. It provides a
1799 callout function that displays information. By default, it shows the start of
1800 the match and the current position in the text. There are some new data escapes
1801 to vary what happens:
1803 \C+ in addition, show current contents of captured substrings
1804 \C- do not supply a callout function
1805 \C!n return 1 when callout number n is reached
1806 \C!n!m return 1 when callout number n is reached for the mth time
1808 30. If pcregrep was called with the -l option and just a single file name, it
1809 output "<stdin>" if a match was found, instead of the file name.
1811 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1812 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1813 pcre_exec(). This saves a malloc/free per call. The default value of
1814 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1815 when configuring.
1817 32. The default maximum size of a compiled pattern is 64K. There have been a
1818 few cases of people hitting this limit. The code now uses macros to handle the
1819 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1820 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1821 configuring. Tests 2 and 5 work only with 2-byte links because they output
1822 debugging information about compiled patterns.
1824 33. Internal code re-arrangements:
1826 (a) Moved the debugging function for printing out a compiled regex into
1827 its own source file (printint.c) and used #include to pull it into
1828 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1829 separate copies.
1831 (b) Defined the list of op-code names for debugging as a macro in
1832 internal.h so that it is next to the definition of the opcodes.
1834 (c) Defined a table of op-code lengths for simpler skipping along compiled
1835 code. This is again a macro in internal.h so that it is next to the
1836 definition of the opcodes.
1838 34. Added support for recursive calls to individual subpatterns, along the
1839 lines of Robin Houston's patch (but implemented somewhat differently).
1841 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1842 allow it to read and process whole directories in Win32. This code was
1843 contributed by Lionel Fourquaux; it has not been tested by me.
1845 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1846 used to name a group. Names consist of alphanumerics and underscores, and must
1847 be unique. Back references use the syntax (?P=name) and recursive calls use
1848 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1849 numbers. The function pcre_fullinfo() can be used after compilation to extract
1850 a name/number map. There are three relevant calls:
1852 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1853 PCRE_INFO_NAMECOUNT yields the number of entries
1854 PCRE_INFO_NAMETABLE yields a pointer to the map.
1856 The map is a vector of fixed-size entries. The size of each entry depends on
1857 the length of the longest name used. The first two bytes of each entry are the
1858 group number, most significant byte first. There follows the corresponding
1859 name, zero terminated. The names are in alphabetical order.
1861 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1862 case instead of 255. Making it the same both with and without UTF-8 support
1863 means that the same test output works with both.
1865 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1866 calling malloc() with a zero argument.
1868 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1869 optimization. I've improved things by keeping a bitmap of backreferences with
1870 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1871 fact referenced, the optimization can be applied. It is unlikely that a
1872 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1873 the match to follow \n) will appear inside brackets with a number greater than
1874 31, but if it does, any back reference > 31 suppresses the optimization.
1876 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1877 of disabling numbered capturing parentheses. Any opening parenthesis that is
1878 not followed by ? behaves as if it were followed by ?: but named parentheses
1879 can still be used for capturing (and they will acquire numbers in the usual
1880 way).
1882 41. Redesigned the return codes from the match() function into yes/no/error so
1883 that errors can be passed back from deep inside the nested calls. A malloc
1884 failure while inside a recursive subpattern call now causes the
1885 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1887 42. It is now possible to set a limit on the number of times the match()
1888 function is called in a call to pcre_exec(). This facility makes it possible to
1889 limit the amount of recursion and backtracking, though not in a directly
1890 obvious way, because the match() function is used in a number of different
1891 circumstances. The count starts from zero for each position in the subject
1892 string (for non-anchored patterns). The default limit is, for compatibility, a
1893 large number, namely 10 000 000. You can change this in two ways:
1895 (a) When configuring PCRE before making, you can use --with-match-limit=n
1896 to set a default value for the compiled library.
1898 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1899 a different value is set. See 45 below.
1901 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1903 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1904 of things that can be changed at compile time. The first argument specifies
1905 what is wanted and the second points to where the information is to be placed.
1906 The current list of available information is:
1910 The output is an integer that is set to one if UTF-8 support is available;
1911 otherwise it is set to zero.
1915 The output is an integer that it set to the value of the code that is used for
1916 newline. It is either LF (10) or CR (13).
1920 The output is an integer that contains the number of bytes used for internal
1921 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1925 The output is an integer that contains the threshold above which the POSIX
1926 interface uses malloc() for output vectors. See item 31 above.
1930 The output is an unsigned integer that contains the default limit of the number
1931 of match() calls in a pcre_exec() execution. See 42 above.
1933 44. pcretest has been upgraded by the addition of the -C option. This causes it
1934 to extract all the available output from the new pcre_config() function, and to
1935 output it. The program then exits immediately.
1937 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1938 order to support additional features. One way would have been to define
1939 pcre_exec2() (for example) with extra arguments, but this would not have been
1940 extensible, and would also have required all calls to the original function to
1941 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1942 is used for passing in "extra" data from pcre_study().
1944 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1945 contains the following fields:
1947 flags a bitmap indicating which of the following fields are set
1948 study_data opaque data from pcre_study()
1949 match_limit a way of specifying a limit on match() calls for a specific
1950 call to pcre_exec()
1951 callout_data data for callouts (see 49 below)
1953 The flag bits are also defined in pcre.h, and are
1959 The pcre_study() function now returns one of these new pcre_extra blocks, with
1960 the actual study data pointed to by the study_data field, and the
1961 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1962 before. That is, this change is entirely upwards-compatible and requires no
1963 change to existing code.
1965 If you want to pass in additional data to pcre_exec(), you can either place it
1966 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1967 block.
1969 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1970 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1971 times with different match limits, until it finds the minimum value needed for
1972 pcre_exec() to complete. The value is then output. This can be instructive; for
1973 most simple matches the number is quite small, but for pathological cases it
1974 gets very large very quickly.
1976 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1977 returns the size of the data block pointed to by the study_data field in a
1978 pcre_extra block, that is, the value that was passed as the argument to
1979 pcre_malloc() when PCRE was getting memory in which to place the information
1980 created by pcre_study(). The fourth argument should point to a size_t variable.
1981 pcretest has been extended so that this information is shown after a successful
1982 pcre_study() call when information about the compiled regex is being displayed.
1984 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1985 because what follows is always an absolute path. (Later: it turns out that this
1986 is more than cosmetic for MinGW, because it doesn't like empty path
1987 components.)
1989 49. Some changes have been made to the callout feature (see 28 above):
1991 (i) A callout function now has three choices for what it returns:
1993 0 => success, carry on matching
1994 > 0 => failure at this point, but backtrack if possible
1995 < 0 => serious error, return this value from pcre_exec()
1997 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1998 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1999 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2000 use by callout functions. It will never be used by PCRE itself.
2002 (ii) The pcre_extra structure (see 45 above) has a void * field called
2003 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2004 pcre_callout_block structure has a field of the same name. The contents of
2005 the field passed in the pcre_extra structure are passed to the callout
2006 function in the corresponding field in the callout block. This makes it
2007 easier to use the same callout-containing regex from multiple threads. For
2008 testing, the pcretest program has a new data escape
2010 \C*n pass the number n (may be negative) as callout_data
2012 If the callout function in pcretest receives a non-zero value as
2013 callout_data, it returns that value.
2015 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2016 there were some redundant $(CFLAGS) in commands that are now specified as
2017 $(LINK), which already includes $(CFLAGS).
2019 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2020 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2021 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2022 one-byte characters throughout. Note that case-insensitive matching applies
2023 only to characters whose values are less than 256. PCRE doesn't support the
2024 notion of cases for higher-valued characters.
2026 (i) A character class whose characters are all within 0-255 is handled as
2027 a bit map, and the map is inverted for negative classes. Previously, a
2028 character > 255 always failed to match such a class; however it should
2029 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2031 (ii) A negated character class with a single character < 255 is coded as
2032 "not this character" (OP_NOT). This wasn't working properly when the test
2033 character was multibyte, either singly or repeated.
2035 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2036 mode, for example: \x{100}{2,3}.
2038 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2039 singly or repeated) now correctly test multibyte characters. However,
2040 PCRE doesn't recognize any characters with values greater than 255 as
2041 digits, spaces, or word characters. Such characters always match \D, \S,
2042 and \W, and never match \d, \s, or \w.
2044 (v) Classes may now contain characters and character ranges with values
2045 greater than 255. For example: [ab\x{100}-\x{400}].
2047 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2048 PCRE in UTF-8 mode.
2050 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2051 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2052 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2053 value.)
2055 53. The single man page has become too large. I have therefore split it up into
2056 a number of separate man pages. These also give rise to individual HTML pages;
2057 these are now put in a separate directory, and there is an index.html page that
2058 lists them all. Some hyperlinking between the pages has been installed.
2060 54. Added convenience functions for handling named capturing parentheses.
2062 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2063 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2064 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2065 are faulted.
2067 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2068 calling configure. These values are used when compiling the dftables.c program
2069 which is run to generate the source of the default character tables. They
2070 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2071 you will need to set these values.
2073 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2076 Version 3.9 02-Jan-02
2077 ---------------------
2079 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2081 2. If --disable-static was given, the building process failed when trying to
2082 build pcretest and pcregrep. (For some reason it was using libtool to compile
2083 them, which is not right, as they aren't part of the library.)
2086 Version 3.8 18-Dec-01
2087 ---------------------
2089 1. The experimental UTF-8 code was completely screwed up. It was packing the
2090 bytes in the wrong order. How dumb can you get?
2093 Version 3.7 29-Oct-01
2094 ---------------------
2096 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2097 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2098 this didn't happen under Solaris 8, where I normally test things.
2100 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2101 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2102 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2103 this fix an hour or so after the initial 3.7 release.)
2106 Version 3.6 23-Oct-01
2107 ---------------------
2109 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2110 offsets passed as NULL with zero offset count.
2112 2. The config.guess and config.sub files had not been updated when I moved to
2113 the latest autoconf.
2116 Version 3.5 15-Aug-01
2117 ---------------------
2119 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2120 had been forgotten.
2122 2. By using declared but undefined structures, we can avoid using "void"
2123 definitions in pcre.h while keeping the internal definitions of the structures
2124 private.
2126 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2127 user point of view, this means that both static and shared libraries are built
2128 by default, but this can be individually controlled. More of the work of
2129 handling this static/shared cases is now inside libtool instead of PCRE's make
2130 file.
2132 4. The pcretest utility is now installed along with pcregrep because it is
2133 useful for users (to test regexs) and by doing this, it automatically gets
2134 relinked by libtool. The documentation has been turned into a man page, so
2135 there are now .1, .txt, and .html versions in /doc.
2137 5. Upgrades to pcregrep:
2138 (i) Added long-form option names like gnu grep.
2139 (ii) Added --help to list all options with an explanatory phrase.
2140 (iii) Added -r, --recursive to recurse into sub-directories.
2141 (iv) Added -f, --file to read patterns from a file.
2143 6. pcre_exec() was referring to its "code" argument before testing that
2144 argument for NULL (and giving an error if it was NULL).
2146 7. Upgraded Makefile.in to allow for compiling in a different directory from
2147 the source directory.
2149 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2150 options bits, the pointer it was passed was to an int instead of to an unsigned
2151 long int. This mattered only on 64-bit systems.
2153 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2154 generated) instead of pcre.in, which it its source. Also made the same change
2155 in several of the .c files.
2157 10. A new release of gcc defines printf() as a macro, which broke pcretest
2158 because it had an ifdef in the middle of a string argument for printf(). Fixed
2159 by using separate calls to printf().
2161 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2162 script, to force use of CR or LF instead of \n in the source. On non-Unix
2163 systems, the value can be set in config.h.
2165 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2166 absolute limit. Changed the text of the error message to make this clear, and
2167 likewise updated the man page.
2169 13. The limit of 99 on the number of capturing subpatterns has been removed.
2170 The new limit is 65535, which I hope will not be a "real" limit.
2173 Version 3.4 22-Aug-00
2174 ---------------------
2176 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2178 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2181 Version 3.3 01-Aug-00
2182 ---------------------
2184 1. If an octal character was given, but the value was greater than \377, it
2185 was not getting masked to the least significant bits, as documented. This could
2186 lead to crashes in some systems.
2188 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2189 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2191 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2192 These just pass their arguments on to (pcre_free)(), but they are provided
2193 because some uses of PCRE bind it to non-C systems that can call its functions,
2194 but cannot call free() or pcre_free() directly.
2196 4. Add "make test" as a synonym for "make check". Corrected some comments in
2197 the Makefile.
2199 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2200 Makefile.
2202 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2203 command called pgrep for grepping around the active processes.
2205 7. Added the beginnings of support for UTF-8 character strings.
2207 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2208 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2209 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2210 out for the ar command.)
2213 Version 3.2 12-May-00
2214 ---------------------
2216 This is purely a bug fixing release.
2218 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2219 of ZA. This was just one example of several cases that could provoke this bug,
2220 which was introduced by change 9 of version 2.00. The code for breaking
2221 infinite loops after an iteration that matches an empty string was't working
2222 correctly.
2224 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2225 when matched against abbab (for example). After matching an empty string, it
2226 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2227 caused it to match further down the string than it should.
2229 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2230 was there because it doesn't seem to be needed, and it causes trouble on some
2231 systems, as it is not a Standard C header. It has been removed.
2233 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2234 were reported on the Macintosh. The changes were from
2236 while ((c = *(++ptr)) != 0 && c != '\n');
2237 to
2238 while ((c = *(++ptr)) != 0 && c != '\n') ;
2240 Totally extraordinary, but if that's what it takes...
2242 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2243 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2244 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2245 assumes the way PCRE uses memmove() (always moving upwards).
2247 6. PCRE is being used in one environment where strchr() is not available. There
2248 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2249 faster code anyway.
2252 Version 3.1 09-Feb-00
2253 ---------------------
2255 The only change in this release is the fixing of some bugs in Makefile.in for
2256 the "install" target:
2258 (1) It was failing to install pcreposix.h.
2260 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2263 Version 3.0 01-Feb-00
2264 ---------------------
2266 1. Add support for the /+ modifier to perltest (to output $` like it does in
2267 pcretest).
2269 2. Add support for the /g modifier to perltest.
2271 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2272 matches null strings.
2274 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2275 pattern. Perl treats empty patterns specially - it reuses the most recent
2276 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2277 effect.
2279 5. The POSIX interface was broken in that it was just handing over the POSIX
2280 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2281 required a bigger vector, with some working space on the end. This means that
2282 the POSIX wrapper now has to get and free some memory, and copy the results.
2284 6. Added some simple autoconf support, placing the test data and the
2285 documentation in separate directories, re-organizing some of the
2286 information files, and making it build pcre-config (a GNU standard). Also added
2287 libtool support for building PCRE as a shared library, which is now the
2288 default.
2290 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2291 09 are not valid octal constants. Single digits will be used for minor values
2292 less than 10.
2294 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2295 existing programs that set these in the POSIX interface can use PCRE without
2296 modification.
2298 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2299 return all that pcre_info() returns, plus additional data. The pcre_info()
2300 function is retained for compatibility, but is considered to be obsolete.
2302 10. Added experimental recursion feature (?R) to handle one common case that
2303 Perl 5.6 will be able to do with (?p{...}).
2305 11. Added support for POSIX character classes like [:alpha:], which Perl is
2306 adopting.
2309 Version 2.08 31-Aug-99
2310 ----------------------
2312 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2313 trying to match at the startoffset position, but instead was moving forward to
2314 the next newline as if a previous match had failed.
2316 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2317 and could get into a loop if a null string was matched other than at the start
2318 of the subject.
2320 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2321 be distinguished at compile time, and for completeness also added PCRE_DATE.
2323 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2324 in GnuWin32 environments.
2327 Version 2.07 29-Jul-99
2328 ----------------------
2330 1. The documentation is now supplied in plain text form and HTML as well as in
2331 the form of man page sources.
2333 2. C++ compilers don't like assigning (void *) values to other pointer types.
2334 In particular this affects malloc(). Although there is no problem in Standard
2335 C, I've put in casts to keep C++ compilers happy.
2337 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2338 should be (const char *).
2340 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2341 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2342 However, I haven't made this a standard facility. The documentation doesn't
2343 mention it, and the Makefile doesn't support it.
2345 5. The Makefile now contains an "install" target, with editable destinations at
2346 the top of the file. The pcretest program is not installed.
2348 6. pgrep -V now gives the PCRE version number and date.
2350 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2351 causing the entire string to be ignored, instead of just the last character.
2353 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2354 non-matching string, it can take a very, very long time, even for strings of
2355 quite modest length, because of the nested recursion. PCRE now does better in
2356 some of these cases. It does this by remembering the last required literal
2357 character in the pattern, and pre-searching the subject to ensure it is present
2358 before running the real match. In other words, it applies a heuristic to detect
2359 some types of certain failure quickly, and in the above example, if presented
2360 with a string that has no trailing " it gives "no match" very quickly.
2362 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2363 other alternatives are tried instead.
2366 Version 2.06 09-Jun-99
2367 ----------------------
2369 1. Change pcretest's output for amount of store used to show just the code
2370 space, because the remainder (the data block) varies in size between 32-bit and
2371 64-bit systems.
2373 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2374 start matching at. This allows lookbehinds to work when searching for multiple
2375 occurrences in a string.
2377 3. Added additional options to pcretest for testing multiple occurrences:
2379 /+ outputs the rest of the string that follows a match
2380 /g loops for multiple occurrences, using the new startoffset argument
2381 /G loops for multiple occurrences by passing an incremented pointer
2383 4. PCRE wasn't doing the "first character" optimization for patterns starting
2384 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2385 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2386 the letter 'x'. On long subject strings, this gives a significant speed-up.
2389 Version 2.05 21-Apr-99
2390 ----------------------
2392 1. Changed the type of magic_number from int to long int so that it works
2393 properly on 16-bit systems.
2395 2. Fixed a bug which caused patterns starting with .* not to work correctly
2396 when the subject string contained newline characters. PCRE was assuming
2397 anchoring for such patterns in all cases, which is not correct because .* will
2398 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2399 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2400 must be retried after every newline in the subject.
2403 Version 2.04 18-Feb-99
2404 ----------------------
2406 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2407 computation of the store needed to hold the pattern was incorrect (too large).
2408 If such patterns were nested a few deep, this could multiply and become a real
2409 problem.
2411 2. Added /M option to pcretest to show the memory requirement of a specific
2412 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2414 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2415 compiled in such a way that the backtracking after subsequent failure was
2416 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2417 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2420 Version 2.03 02-Feb-99
2421 ----------------------
2423 1. Fixed typo and small mistake in man page.
2425 2. Added 4th condition (GPL supersedes if conflict) and created separate
2426 LICENCE file containing the conditions.
2428 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2429 Perl, that is the internal \ allows the delimiter to be included in the
2430 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2431 the final delimiter, add \ to the end of the pattern (to test the error).
2433 4. Added the convenience functions for extracting substrings after a successful
2434 match. Updated pcretest to make it able to test these functions.
2437 Version 2.02 14-Jan-99
2438 ----------------------
2440 1. Initialized the working variables associated with each extraction so that
2441 their saving and restoring doesn't refer to uninitialized store.
2443 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2444 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2445 fix the problem.
2447 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2448 calls, and wasn't printing the correct value for compiling calls. Increased the
2449 default value of LOOPREPEAT, and the number of significant figures in the
2450 times.
2452 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2454 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2455 a building problem on Windows NT with a FAT file system.
2458 Version 2.01 21-Oct-98
2459 ----------------------
2461 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2462 to character tables built by pcre_maketables() in the current locale. If NULL
2463 is passed, the default tables are used.
2466 Version 2.00 24-Sep-98
2467 ----------------------
2469 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2470 it any more.
2472 2. Allow quantification of (?>) groups, and make it work correctly.
2474 3. The first character computation wasn't working for (?>) groups.
2476 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2477 end of the subject) and add 5.005's \z, which really does match only at the
2478 very end of the subject.
2480 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2482 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2483 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2484 localized options. All options to pcre_study() were also removed.
2486 7. Add other new features from 5.005:
2488 $(?<= positive lookbehind
2489 $(?<! negative lookbehind
2490 (?imsx-imsx) added the unsetting capability
2491 such a setting is global if at outer level; local otherwise
2492 (?imsx-imsx:) non-capturing groups with option setting
2493 (?(cond)re|re) conditional pattern matching
2495 A backreference to itself in a repeated group matches the previous
2496 captured string.
2498 8. General tidying up of studying (both automatic and via "study")
2499 consequential on the addition of new assertions.
2501 9. As in 5.005, unlimited repeated groups that could match an empty substring
2502 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2503 runtime if any iteration does actually match an empty substring.
2505 10. Include the RunTest script in the distribution.
2507 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2508 discrepancies, some of which were old and were also with respect to 5.004. They
2509 have now been fixed.
2512 Version 1.09 28-Apr-98
2513 ----------------------
2515 1. A negated single character class followed by a quantifier with a minimum
2516 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2517 program crashes, or just wrong answers. This did not apply to negated classes
2518 containing more than one character, or to minima other than one.
2521 Version 1.08 27-Mar-98
2522 ----------------------
2524 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2526 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2527 latter must appear before anything that relies on it in the pattern.
2530 Version 1.07 16-Feb-98
2531 ----------------------
2533 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2534 repeat of a potentially empty string).
2537 Version 1.06 23-Jan-98
2538 ----------------------
2540 1. Added Markus Oberhumer's little patches for C++.
2542 2. Literal strings longer than 255 characters were broken.
2545 Version 1.05 23-Dec-97
2546 ----------------------
2548 1. Negated character classes containing more than one character were failing if
2549 PCRE_CASELESS was set at run time.
2552 Version 1.04 19-Dec-97
2553 ----------------------
2555 1. Corrected the man page, where some "const" qualifiers had been omitted.
2557 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2558 input syntax.
2560 3. Fixed memory leak which occurred when a regex with back references was
2561 matched with an offsets vector that wasn't big enough. The temporary memory
2562 that is used in this case wasn't being freed if the match failed.
2564 4. Tidied pcretest to ensure it frees memory that it gets.
2566 5. Temporary memory was being obtained in the case where the passed offsets
2567 vector was exactly big enough.
2569 6. Corrected definition of offsetof() from change 5 below.
2571 7. I had screwed up change 6 below and broken the rules for the use of
2572 setjmp(). Now fixed.
2575 Version 1.03 18-Dec-97
2576 ----------------------
2578 1. A erroneous regex with a missing opening parenthesis was correctly
2579 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2580 on some systems.
2582 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2583 it was reported that one broken compiler failed on the former because "code" is
2584 also an independent variable.
2586 3. The erroneous regex a[]b caused an array overrun reference.
2588 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2589 fail on data ending with that character. (It was going on too far, and checking
2590 the next character, typically a binary zero.) This was specific to the
2591 optimized code for single-character negative classes.
2593 5. Added a contributed patch from the TIN world which does the following:
2595 + Add an undef for memmove, in case the the system defines a macro for it.
2597 + Add a definition of offsetof(), in case there isn't one. (I don't know
2598 the reason behind this - offsetof() is part of the ANSI standard - but
2599 it does no harm).
2601 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2602 most of the places where whitespace preceded '#'. I have given up and
2603 allowed the remaining 2 cases to be at the margin.
2605 + Rename some variables in pcre to eliminate shadowing. This seems very
2606 pedantic, but does no harm, of course.
2608 6. Moved the call to setjmp() into its own function, to get rid of warnings
2609 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2611 7. Constructs such as \d{8,} were compiling into the equivalent of
2612 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2613 outcome, but in this particular case used more store than had been allocated,
2614 which caused the bug to be discovered because it threw up an internal error.
2616 8. The debugging code in both pcre and pcretest for outputting the compiled
2617 form of a regex was going wrong in the case of back references followed by
2618 curly-bracketed repeats.
2621 Version 1.02 12-Dec-97
2622 ----------------------
2624 1. Typos in pcre.3 and comments in the source fixed.
2626 2. Applied a contributed patch to get rid of places where it used to remove
2627 'const' from variables, and fixed some signed/unsigned and uninitialized
2628 variable warnings.
2630 3. Added the "runtest" target to Makefile.
2632 4. Set default compiler flag to -O2 rather than just -O.
2635 Version 1.01 19-Nov-97
2636 ----------------------
2638 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2639 like /([ab]*)*/, that is, for classes with more than one character in them.
2641 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2642 as /((?>a*))*/ (a PCRE_EXTRA facility).
2645 Version 1.00 18-Nov-97
2646 ----------------------
2648 1. Added compile-time macros to support systems such as SunOS4 which don't have
2649 memmove() or strerror() but have other things that can be used instead.
2651 2. Arranged that "make clean" removes the executables.
2654 Version 0.99 27-Oct-97
2655 ----------------------
2657 1. Fixed bug in code for optimizing classes with only one character. It was
2658 initializing a 32-byte map regardless, which could cause it to run off the end
2659 of the memory it had got.
2661 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2664 Version 0.98 22-Oct-97
2665 ----------------------
2667 1. Fixed bug in code for handling temporary memory usage when there are more
2668 back references than supplied space in the ovector. This could cause segfaults.
2671 Version 0.97 21-Oct-97
2672 ----------------------
2674 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2676 2. Optimized negated single characters not to use a bit map.
2678 3. Brought error texts together as macro definitions; clarified some of them;
2679 fixed one that was wrong - it said "range out of order" when it meant "invalid
2680 escape sequence".
2682 4. Changed some char * arguments to const char *.
2684 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2686 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2687 pcretest.
2690 Version 0.96 16-Oct-97
2691 ----------------------
2693 1. Added a simple "pgrep" utility to the distribution.
2695 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2696 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2697 where "ddd" means "one or more decimal digits".
2699 3. Fixed serious bug. If a pattern had a back reference, but the call to
2700 pcre_exec() didn't supply a large enough ovector to record the related
2701 identifying subpattern, the match always failed. PCRE now remembers the number
2702 of the largest back reference, and gets some temporary memory in which to save
2703 the offsets during matching if necessary, in order to ensure that
2704 backreferences always work.
2706 4. Increased the compatibility with Perl in a number of ways:
2708 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2709 to request this handling. The option can be set at compile or exec time.
2711 (b) $ matches before a terminating newline by default; an option
2712 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2713 mode). The option can be set at compile or exec time.
2715 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2716 the same as Perl's. If the decimal number it represents is less than 10
2717 or there aren't that many previous left capturing parentheses, an octal
2718 escape is read. Inside a character class, it's always an octal escape,
2719 even if it is a single digit.
2721 (d) An escaped but undefined alphabetic character is taken as a literal,
2722 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2723 escapes.
2725 (e) {0} is now permitted. (The previous item is removed from the compiled
2726 pattern).
2728 5. Changed all the names of code files so that the basic parts are no longer
2729 than 10 characters, and abolished the teeny "globals.c" file.
2731 6. Changed the handling of character classes; they are now done with a 32-byte
2732 bit map always.
2734 7. Added the -d and /D options to pcretest to make it possible to look at the
2735 internals of compilation without having to recompile pcre.
2738 Version 0.95 23-Sep-97
2739 ----------------------
2741 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2742 \x20 at the start of a run of normal characters. These were being treated as
2743 real characters, instead of the source characters being re-checked.
2746 Version 0.94 18-Sep-97
2747 ----------------------
2749 1. The functions are now thread-safe, with the caveat that the global variables
2750 containing pointers to malloc() and free() or alternative functions are the
2751 same for all threads.
2753 2. Get pcre_study() to generate a bitmap of initial characters for non-
2754 anchored patterns when this is possible, and use it if passed to pcre_exec().
2757 Version 0.93 15-Sep-97
2758 ----------------------
2760 1. /(b)|(:+)/ was computing an incorrect first character.
2762 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2763 but not actually doing anything yet.
2765 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2766 as Perl does (e.g. [-az] or [az-]).
2768 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2769 all possible positions.
2771 5. Split up into different modules to avoid including unneeded functions in a
2772 compiled binary. However, compile and exec are still in one module. The "study"
2773 function is split off.
2775 6. The character tables are now in a separate module whose source is generated
2776 by an auxiliary program - but can then be edited by hand if required. There are
2777 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2778 toupper() in the code.
2780 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2781 make them global. Abolish the function for setting them, as the caller can now
2782 set them directly.
2785 Version 0.92 11-Sep-97
2786 ----------------------
2788 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2789 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2791 2. Caseless matching was not working in character classes if the characters in
2792 the pattern were in upper case.
2794 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2796 4. Make PCRE_ANCHORED public and accept as a compile option.
2798 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2799 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2800 pass them.
2802 6. Give an error if bad option bits passed at compile or run time.
2804 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2805 pcretest to cause it to pass that flag.
2807 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2808 options, and the first character, if set.
2810 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2813 Version 0.91 10-Sep-97
2814 ----------------------
2816 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2817 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2819 2. PCRE was looping on encountering an indefinitely repeated back reference to
2820 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2821 Perl does - treats the match as successful.
2823 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5