ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 111 - (show annotations)
Thu Mar 8 16:53:09 2007 UTC (14 years, 7 months ago) by ph10
File size: 122681 byte(s)
Create the PrepareRelease script to process the documentation and create the 
.generic files for distribution, also to remove trailing spaces. Update a lot 
more of the build-time documentation. Arrange for PrepareRelease and its 
sub-scripts to be distributed.
1 ChangeLog for PCRE
2 ------------------
4 Version 7.1 05-Mar-07
5 ---------------------
7 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
8 that is more "standard", making use of automake and other autotools. There
9 is some re-arrangement of the files and adjustment of comments consequent
10 on this.
12 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
13 for recursive directory scanning broke on some systems because the files
14 are not scanned in any specific order and on different systems the order
15 was different. A call to "sort" has been inserted into RunGrepTest for the
16 approprate test as a short-term fix. In the longer term there may be an
17 alternative.
19 3. I had an email from Eric Raymond about problems translating some of PCRE's
20 man pages to HTML (despite the fact that I distribute HTML pages, some
21 people do their own conversions for various reasons). The problems
22 concerned the use of low-level troff macros .br and .in. I have therefore
23 removed all such uses from the man pages (some were redundant, some could
24 be replaced by .nf/.fi pairs). The maintain/132html script that I use to
25 generate HTML has been updated to handle .nf/.fi and to complain if it
26 encounters .br or .in.
28 4. Updated comments in configure.ac that get placed in config.h.in and also
29 arranged for config.h to be included in the distribution, with the name
30 config.h.generic, for the benefit of those who have to compile without
31 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
33 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
34 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
35 makevp.bat for the current PCRE, using the additional files !compile.txt,
36 !linklib.txt, and pcregexp.pas.
38 6. A Windows user reported a minor discrepancy with test 2, which turned out
39 to be caused by a trailing space on an input line that had got lost in his
40 copy. The trailing space was an accident, so I've just removed it.
42 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
43 that is needed.
45 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
46 as "const" (a) because they are and (b) because it helps the PHP
47 maintainers who have recently made a script to detect big data structures
48 in the php code that should be moved to the .rodata section. I remembered
49 to update Builducptable as well, so it won't revert if ucptable.h is ever
50 re-created.
52 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
53 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
54 order to be able to cut out the UTF-8 tables in the latter when UTF-8
55 support is not required. This saves 1.5-2K of code, which is important in
56 some applications.
58 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
59 so as not to refer to the tables, even though these functions will never be
60 called when UTF-8 support is disabled. Otherwise there are problems with a
61 shared library.
63 10. The emulated memmove() function in pcre_internal.h (provided for those
64 environments that have neither memmove() nor bcopy()) was defining its
65 arguments as char * instead of void *.
67 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
68 and Detrail to create the HTML documentation, the .txt form of the man
69 pages, and it removes trailing spaces from listed files. It also creates
70 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
71 case, it wraps all the #defines with #ifndefs. This script should be run
72 before "make dist".
75 Version 7.0 19-Dec-06
76 ---------------------
78 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
79 moving to gcc 4.1.1.
81 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
82 sys/time.h, which is documented as needed for this function. It doesn't
83 seem to matter on Linux, but it showed up on some releases of OS X.
85 3. It seems that there are systems where bytes whose values are greater than
86 127 match isprint() in the "C" locale. The "C" locale should be the
87 default when a C program starts up. In most systems, only ASCII printing
88 characters match isprint(). This difference caused the output from pcretest
89 to vary, making some of the tests fail. I have changed pcretest so that:
91 (a) When it is outputting text in the compiled version of a pattern, bytes
92 other than 32-126 are always shown as hex escapes.
94 (b) When it is outputting text that is a matched part of a subject string,
95 it does the same, unless a different locale has been set for the match
96 (using the /L modifier). In this case, it uses isprint() to decide.
98 4. Fixed a major bug that caused incorrect computation of the amount of memory
99 required for a compiled pattern when options that changed within the
100 pattern affected the logic of the preliminary scan that determines the
101 length. The relevant options are -x, and -i in UTF-8 mode. The result was
102 that the computed length was too small. The symptoms of this bug were
103 either the PCRE error "internal error: code overflow" from pcre_compile(),
104 or a glibc crash with a message such as "pcretest: free(): invalid next
105 size (fast)". Examples of patterns that provoked this bug (shown in
106 pcretest format) are:
108 /(?-x: )/x
109 /(?x)(?-x: \s*#\s*)/
110 /((?i)[\x{c0}])/8
111 /(?i:[\x{c0}])/8
113 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
114 is now done differently.
116 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
117 wrapper classes; (b) implement a new function in the C++ scanner that is
118 more efficient than the old way of doing things because it avoids levels of
119 recursion in the regex matching; (c) add a paragraph to the documentation
120 for the FullMatch() function.
122 6. The escape sequence \n was being treated as whatever was defined as
123 "newline". Not only was this contrary to the documentation, which states
124 that \n is character 10 (hex 0A), but it also went horribly wrong when
125 "newline" was defined as CRLF. This has been fixed.
127 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
128 was being set to -1 for the "end of line" case (supposedly a value that no
129 character can have). Though this value is never used (the check for end of
130 line is "zero bytes in current character"), it caused compiler complaints.
131 I've changed it to 0xffffffff.
133 8. In pcre_version.c, the version string was being built by a sequence of
134 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
135 string (as it is for production releases) called a macro with an empty
136 argument. The C standard says the result of this is undefined. The gcc
137 compiler treats it as an empty string (which was what was wanted) but it is
138 reported that Visual C gives an error. The source has been hacked around to
139 avoid this problem.
141 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
142 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
143 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
144 of them did).
146 10. Originally, pcretest opened its input and output without "b"; then I was
147 told that "b" was needed in some environments, so it was added for release
148 5.0 to both the input and output. (It makes no difference on Unix-like
149 systems.) Later I was told that it is wrong for the input on Windows. I've
150 now abstracted the modes into two macros, to make it easier to fiddle with
151 them, and removed "b" from the input mode under Windows.
153 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
155 12. Added -help and --help to pcretest as an official way of being reminded
156 of the options.
158 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
159 and pcrecpp.cc because they annoy compilers at high warning levels.
161 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
163 15. Fixed an occurrence of == in configure.ac that should have been = (shell
164 scripts are not C programs :-) and which was not noticed because it works
165 on Linux.
167 16. pcretest is supposed to handle any length of pattern and data line (as one
168 line or as a continued sequence of lines) by extending its input buffer if
169 necessary. This feature was broken for very long pattern lines, leading to
170 a string of junk being passed to pcre_compile() if the pattern was longer
171 than about 50K.
173 17. I have done a major re-factoring of the way pcre_compile() computes the
174 amount of memory needed for a compiled pattern. Previously, there was code
175 that made a preliminary scan of the pattern in order to do this. That was
176 OK when PCRE was new, but as the facilities have expanded, it has become
177 harder and harder to keep it in step with the real compile phase, and there
178 have been a number of bugs (see for example, 4 above). I have now found a
179 cunning way of running the real compile function in a "fake" mode that
180 enables it to compute how much memory it would need, while actually only
181 ever using a few hundred bytes of working memory and without too many
182 tests of the mode. This should make future maintenance and development
183 easier. A side effect of this work is that the limit of 200 on the nesting
184 depth of parentheses has been removed (though this was never a serious
185 limitation, I suspect). However, there is a downside: pcre_compile() now
186 runs more slowly than before (30% or more, depending on the pattern). I
187 hope this isn't a big issue. There is no effect on runtime performance.
189 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
190 newline (only possible for the last line of a file) and it was a
191 pattern that set a locale (followed by /Lsomething), pcretest crashed.
193 19. Added additional timing features to pcretest. (1) The -tm option now times
194 matching only, not compiling. (2) Both -t and -tm can be followed, as a
195 separate command line item, by a number that specifies the number of
196 repeats to use when timing. The default is 50000; this gives better
197 precision, but takes uncomfortably long for very large patterns.
199 20. Extended pcre_study() to be more clever in cases where a branch of a
200 subpattern has no definite first character. For example, (a*|b*)[cd] would
201 previously give no result from pcre_study(). Now it recognizes that the
202 first character must be a, b, c, or d.
204 21. There was an incorrect error "recursive call could loop indefinitely" if
205 a subpattern (or the entire pattern) that was being tested for matching an
206 empty string contained only one non-empty item after a nested subpattern.
207 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
208 incorrectly, because the \d was being skipped in the check.
210 22. The pcretest program now has a new pattern option /B and a command line
211 option -b, which is equivalent to adding /B to every pattern. This causes
212 it to show the compiled bytecode, without the additional information that
213 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
214 is the same as /B/I).
216 23. A new optimization is now able automatically to treat some sequences such
217 as a*b as a*+b. More specifically, if something simple (such as a character
218 or a simple class like \d) has an unlimited quantifier, and is followed by
219 something that cannot possibly match the quantified thing, the quantifier
220 is automatically "possessified".
222 24. A recursive reference to a subpattern whose number was greater than 39
223 went wrong under certain circumstances in UTF-8 mode. This bug could also
224 have affected the operation of pcre_study().
226 25. Realized that a little bit of performance could be had by replacing
227 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
229 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
231 27. Possessive quantifiers such as a++ were previously implemented by turning
232 them into atomic groups such as ($>a+). Now they have their own opcodes,
233 which improves performance. This includes the automatically created ones
234 from 23 above.
236 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
237 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
238 the first matched character to be a colon. This applied both to named and
239 numbered groups.
241 29. The ucpinternal.h header file was missing its idempotency #ifdef.
243 30. I was sent a "project" file called libpcre.a.dev which I understand makes
244 building PCRE on Windows easier, so I have included it in the distribution.
246 31. There is now a check in pcretest against a ridiculously large number being
247 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
248 loop, the loop is abandoned.
250 32. Forward references to subpatterns in conditions such as (?(2)...) where
251 subpattern 2 is defined later cause pcre_compile() to search forwards in
252 the pattern for the relevant set of parentheses. This search went wrong
253 when there were unescaped parentheses in a character class, parentheses
254 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
256 33. "Subroutine" calls and backreferences were previously restricted to
257 referencing subpatterns earlier in the regex. This restriction has now
258 been removed.
260 34. Added a number of extra features that are going to be in Perl 5.10. On the
261 whole, these are just syntactic alternatives for features that PCRE had
262 previously implemented using the Python syntax or my own invention. The
263 other formats are all retained for compatibility.
265 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
266 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
267 also .NET compatible.
269 (b) A recursion or subroutine call to a named group can now be defined as
270 (?&name) as well as (?P>name).
272 (c) A backreference to a named group can now be defined as \k<name> or
273 \k'name' as well as (?P=name). The new forms, as well as being in Perl
274 5.10, are also .NET compatible.
276 (d) A conditional reference to a named group can now use the syntax
277 (?(<name>) or (?('name') as well as (?(name).
279 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
280 groups (named and numbered) that are never evaluated inline, but can be
281 called as "subroutines" from elsewhere. In effect, the DEFINE condition
282 is always false. There may be only one alternative in such a group.
284 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
285 as the simple (?(R). The condition is true only if the most recent
286 recursion is that of the given number or name. It does not search out
287 through the entire recursion stack.
289 (g) The escape \gN or \g{N} has been added, where N is a positive or
290 negative number, specifying an absolute or relative reference.
292 35. Tidied to get rid of some further signed/unsigned compiler warnings and
293 some "unreachable code" warnings.
295 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
296 things, this adds five new scripts.
298 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
299 There were also incompatibilities regarding the handling of \Q..\E inside
300 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
301 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
303 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
304 matches an empty string, and forcibly breaks the loop. There were bugs in
305 this code in non-simple cases. For a pattern such as ^(a()*)* matched
306 against aaaa the result was just "a" rather than "aaaa", for example. Two
307 separate and independent bugs (that affected different cases) have been
308 fixed.
310 39. Refactored the code to abolish the use of different opcodes for small
311 capturing bracket numbers. This is a tidy that I avoided doing when I
312 removed the limit on the number of capturing brackets for 3.5 back in 2001.
313 The new approach is not only tidier, it makes it possible to reduce the
314 memory needed to fix the previous bug (38).
316 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
317 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
318 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
319 mode.
321 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
322 report.
324 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
325 copy construction and assignment in the C++ wrapper.
327 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
328 couple of bugs that could have given wrong results in the "--newline=crlf"
329 case.
331 44. Added a number of casts and did some reorganization of signed/unsigned int
332 variables following suggestions from Dair Grant. Also renamed the variable
333 "this" as "item" because it is a C++ keyword.
335 45. Arranged for dftables to add
337 #include "pcre_internal.h"
339 to pcre_chartables.c because without it, gcc 4.x may remove the array
340 definition from the final binary if PCRE is built into a static library and
341 dead code stripping is activated.
343 46. For an unanchored pattern, if a match attempt fails at the start of a
344 newline sequence, and the newline setting is CRLF or ANY, and the next two
345 characters are CRLF, advance by two characters instead of one.
348 Version 6.7 04-Jul-06
349 ---------------------
351 1. In order to handle tests when input lines are enormously long, pcretest has
352 been re-factored so that it automatically extends its buffers when
353 necessary. The code is crude, but this _is_ just a test program. The
354 default size has been increased from 32K to 50K.
356 2. The code in pcre_study() was using the value of the re argument before
357 testing it for NULL. (Of course, in any sensible call of the function, it
358 won't be NULL.)
360 3. The memmove() emulation function in pcre_internal.h, which is used on
361 systems that lack both memmove() and bcopy() - that is, hardly ever -
362 was missing a "static" storage class specifier.
364 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
365 containing an extended class (one that cannot be represented by a bitmap
366 because it contains high-valued characters or Unicode property items, e.g.
367 [\pZ]). Almost always one would set UTF-8 mode when processing such a
368 pattern, but PCRE should not loop if you do not (it no longer does).
369 [Detail: two cases were found: (a) a repeated subpattern containing an
370 extended class; (b) a recursive reference to a subpattern that followed a
371 previous extended class. It wasn't skipping over the extended class
372 correctly when UTF-8 mode was not set.]
374 5. A negated single-character class was not being recognized as fixed-length
375 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
376 compile error "lookbehind assertion is not fixed length".
378 6. The RunPerlTest auxiliary script was showing an unexpected difference
379 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
380 write a Perl script that can interpret lines of an input file either as
381 byte characters or as UTF-8, which is what "perltest" was being required to
382 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
383 can't do is switch easily at run time between having the "use utf8;" pragma
384 or not. In the end, I fudged it by using the RunPerlTest script to insert
385 "use utf8;" explicitly for the UTF-8 tests.
387 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
388 the end of the subject string, contrary to the documentation and to what
389 Perl does. This was true of both matching functions. Now it matches only at
390 the start of the subject and immediately after *internal* newlines.
392 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
393 a pointer to an int instead of a pointer to an unsigned long int. This
394 caused problems on 64-bit systems.
396 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
397 instance of the 'standard' template library not being so standard".
399 10. There was no check on the number of named subpatterns nor the maximum
400 length of a subpattern name. The product of these values is used to compute
401 the size of the memory block for a compiled pattern. By supplying a very
402 long subpattern name and a large number of named subpatterns, the size
403 computation could be caused to overflow. This is now prevented by limiting
404 the length of names to 32 characters, and the number of named subpatterns
405 to 10,000.
407 11. Subpatterns that are repeated with specific counts have to be replicated in
408 the compiled pattern. The size of memory for this was computed from the
409 length of the subpattern and the repeat count. The latter is limited to
410 65535, but there was no limit on the former, meaning that integer overflow
411 could in principle occur. The compiled length of a repeated subpattern is
412 now limited to 30,000 bytes in order to prevent this.
414 12. Added the optional facility to have named substrings with the same name.
416 13. Added the ability to use a named substring as a condition, using the
417 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
418 are numbers (not recommended). Forward references are permitted.
420 14. Added forward references in named backreferences (if you see what I mean).
422 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
423 pattern could run off the end of the subject. For example, the pattern
424 "(?s)(.{1,5})"8 did this with the subject "ab".
426 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
427 PCRE_CASELESS was set when matching characters that were quantified with ?
428 or *.
430 17. A character class other than a single negated character that had a minimum
431 but no maximum quantifier - for example [ab]{6,} - was not handled
432 correctly by pce_dfa_exec(). It would match only one character.
434 18. A valid (though odd) pattern that looked like a POSIX character
435 class but used an invalid character after [ (for example [[,abc,]]) caused
436 pcre_compile() to give the error "Failed: internal error: code overflow" or
437 in some cases to crash with a glibc free() error. This could even happen if
438 the pattern terminated after [[ but there just happened to be a sequence of
439 letters, a binary zero, and a closing ] in the memory that followed.
441 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
442 over the years. Originally (before any Unicode support), just the bottom 8
443 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
444 output from "man perlunicode" includes this:
446 The regular expression compiler produces polymorphic opcodes. That
447 is, the pattern adapts to the data and automatically switches to
448 the Unicode character scheme when presented with Unicode data--or
449 instead uses a traditional byte scheme when presented with byte
450 data.
452 Sadly, a wide octal escape does not cause a switch, and in a string with
453 no other multibyte characters, these octal escapes are treated as before.
454 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
455 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
456 Unicode string.
458 I have not perpetrated such confusion in PCRE. Up till now, it took just
459 the bottom 8 bits, as in old Perl. I have now made octal escapes with
460 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
461 translate to the appropriate multibyte character.
463 29. Applied some refactoring to reduce the number of warnings from Microsoft
464 and Borland compilers. This has included removing the fudge introduced
465 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
466 a warning about an unused variable.
468 21. PCRE has not included VT (character 0x0b) in the set of whitespace
469 characters since release 4.0, because Perl (from release 5.004) does not.
470 [Or at least, is documented not to: some releases seem to be in conflict
471 with the documentation.] However, when a pattern was studied with
472 pcre_study() and all its branches started with \s, PCRE still included VT
473 as a possible starting character. Of course, this did no harm; it just
474 caused an unnecessary match attempt.
476 22. Removed a now-redundant internal flag bit that recorded the fact that case
477 dependency changed within the pattern. This was once needed for "required
478 byte" processing, but is no longer used. This recovers a now-scarce options
479 bit. Also moved the least significant internal flag bit to the most-
480 significant bit of the word, which was not previously used (hangover from
481 the days when it was an int rather than a uint) to free up another bit for
482 the future.
484 23. Added support for CRLF line endings as well as CR and LF. As well as the
485 default being selectable at build time, it can now be changed at runtime
486 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
487 specify that it is scanning data with non-default line endings.
489 24. Changed the definition of CXXLINK to make it agree with the definition of
490 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
492 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
493 recursions. This makes a big different to stack usage for some patterns.
495 26. If a subpattern containing a named recursion or subroutine reference such
496 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
497 the space required for the compiled pattern went wrong and gave too small a
498 value. Depending on the environment, this could lead to "Failed: internal
499 error: code overflow at offset 49" or "glibc detected double free or
500 corruption" errors.
502 27. Applied patches from Google (a) to support the new newline modes and (b) to
503 advance over multibyte UTF-8 characters in GlobalReplace.
505 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
506 difference for some implementation of PCRE in some Windows version.
508 29. Added some extra testing facilities to pcretest:
510 \q<number> in a data line sets the "match limit" value
511 \Q<number> in a data line sets the "match recursion limt" value
512 -S <number> sets the stack size, where <number> is in megabytes
514 The -S option isn't available for Windows.
517 Version 6.6 06-Feb-06
518 ---------------------
520 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
521 in pcreposix.h. I have copied the definition from pcre.h.
523 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
524 because pcre.h is no longer a built file.
526 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
527 not normally included in the compiled code.
530 Version 6.5 01-Feb-06
531 ---------------------
533 1. When using the partial match feature with pcre_dfa_exec(), it was not
534 anchoring the second and subsequent partial matches at the new starting
535 point. This could lead to incorrect results. For example, with the pattern
536 /1234/, partially matching against "123" and then "a4" gave a match.
538 2. Changes to pcregrep:
540 (a) All non-match returns from pcre_exec() were being treated as failures
541 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
542 error message is output. Some extra information is given for the
544 probably the only errors that are likely to be caused by users (by
545 specifying a regex that has nested indefinite repeats, for instance).
546 If there are more than 20 of these errors, pcregrep is abandoned.
548 (b) A binary zero was treated as data while matching, but terminated the
549 output line if it was written out. This has been fixed: binary zeroes
550 are now no different to any other data bytes.
552 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
553 used to set a locale for matching. The --locale=xxxx long option has
554 been added (no short equivalent) to specify a locale explicitly on the
555 pcregrep command, overriding the environment variables.
557 (d) When -B was used with -n, some line numbers in the output were one less
558 than they should have been.
560 (e) Added the -o (--only-matching) option.
562 (f) If -A or -C was used with -c (count only), some lines of context were
563 accidentally printed for the final match.
565 (g) Added the -H (--with-filename) option.
567 (h) The combination of options -rh failed to suppress file names for files
568 that were found from directory arguments.
570 (i) Added the -D (--devices) and -d (--directories) options.
572 (j) Added the -F (--fixed-strings) option.
574 (k) Allow "-" to be used as a file name for -f as well as for a data file.
576 (l) Added the --colo(u)r option.
578 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
579 is not present by default.
581 3. A nasty bug was discovered in the handling of recursive patterns, that is,
582 items such as (?R) or (?1), when the recursion could match a number of
583 alternatives. If it matched one of the alternatives, but subsequently,
584 outside the recursion, there was a failure, the code tried to back up into
585 the recursion. However, because of the way PCRE is implemented, this is not
586 possible, and the result was an incorrect result from the match.
588 In order to prevent this happening, the specification of recursion has
589 been changed so that all such subpatterns are automatically treated as
590 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
592 4. I had overlooked the fact that, in some locales, there are characters for
593 which isalpha() is true but neither isupper() nor islower() are true. In
594 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
595 and ordfeminine) are like this. This affected the treatment of \w and \W
596 when they appeared in character classes, but not when they appeared outside
597 a character class. The bit map for "word" characters is now created
598 separately from the results of isalnum() instead of just taking it from the
599 upper, lower, and digit maps. (Plus the underscore character, of course.)
601 5. The above bug also affected the handling of POSIX character classes such as
602 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
603 permanent tables. Instead, the bit maps for such a class were previously
604 created as the appropriate unions of the upper, lower, and digit bitmaps.
605 Now they are created by subtraction from the [[:word:]] class, which has
606 its own bitmap.
608 6. The [[:blank:]] character class matches horizontal, but not vertical space.
609 It is created by subtracting the vertical space characters (\x09, \x0a,
610 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
611 subtraction was done in the overall bitmap for a character class, meaning
612 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
613 be recognized. This bug has been fixed.
615 7. Patches from the folks at Google:
617 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
618 real life, but is still worth protecting against".
620 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
621 regular expressions".
623 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
624 have it.
626 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
627 "configure" and the latter not, in order to fix a problem somebody had
628 with compiling the Arg class on HP-UX.
630 (e) Improve the error-handling of the C++ wrapper a little bit.
632 (f) New tests for checking recursion limiting.
634 8. The pcre_memmove() function, which is used only if the environment does not
635 have a standard memmove() function (and is therefore rarely compiled),
636 contained two bugs: (a) use of int instead of size_t, and (b) it was not
637 returning a result (though PCRE never actually uses the result).
639 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
640 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
641 returned instead of calling malloc() with an overflowing number that would
642 most likely cause subsequent chaos.
644 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
646 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
647 with this option is matched, the nmatch and pmatch options of regexec() are
648 ignored.
650 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
651 provided in case anyone wants to the the POSIX interface with UTF-8
652 strings.
654 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
655 C++ linking (needed for some HP-UX environments).
657 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
658 (unused parameter) and in the pcre_printint() function (omitted "default"
659 switch label when the default is to do nothing).
661 15. Added some code to make it possible, when PCRE is compiled as a C++
662 library, to replace subject pointers for pcre_exec() with a smart pointer
663 class, thus making it possible to process discontinuous strings.
665 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
666 much the same function. They were added by different people who were trying
667 to make PCRE easy to compile on non-Unix systems. It has been suggested
668 that PCRE_EXPORT be abolished now that there is more automatic apparatus
669 for compiling on Windows systems. I have therefore replaced it with
670 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
671 defaults to "extern" for C or "extern C" for C++, which works fine on
672 Unix-like systems. It is now possible to override the value of PCRE_DATA_
673 SCOPE with something explicit in config.h. In addition:
675 (a) pcreposix.h still had just "extern" instead of either of these macros;
676 I have replaced it with PCRE_DATA_SCOPE.
678 (b) Functions such as _pcre_xclass(), which are internal to the library,
679 but external in the C sense, all had PCRE_EXPORT in their definitions.
680 This is apparently wrong for the Windows case, so I have removed it.
681 (It makes no difference on Unix-like systems.)
683 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
684 of recursive calls to match(). This is different to MATCH_LIMIT because
685 that limits the total number of calls to match(), not all of which increase
686 the depth of recursion. Limiting the recursion depth limits the amount of
687 stack (or heap if NO_RECURSE is set) that is used. The default can be set
688 when PCRE is compiled, and changed at run time. A patch from Google adds
689 this functionality to the C++ interface.
691 18. Changes to the handling of Unicode character properties:
693 (a) Updated the table to Unicode 4.1.0.
695 (b) Recognize characters that are not in the table as "Cn" (undefined).
697 (c) I revised the way the table is implemented to a much improved format
698 which includes recognition of ranges. It now supports the ranges that
699 are defined in UnicodeData.txt, and it also amalgamates other
700 characters into ranges. This has reduced the number of entries in the
701 table from around 16,000 to around 3,000, thus reducing its size
702 considerably. I realized I did not need to use a tree structure after
703 all - a binary chop search is just as efficient. Having reduced the
704 number of entries, I extended their size from 6 bytes to 8 bytes to
705 allow for more data.
707 (d) Added support for Unicode script names via properties such as \p{Han}.
709 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
710 matching that character.
712 20. When matching a repeated Unicode property with a minimum greater than zero,
713 (for example \pL{2,}), PCRE could look past the end of the subject if it
714 reached it while seeking the minimum number of characters. This could
715 happen only if some of the characters were more than one byte long, because
716 there is a check for at least the minimum number of bytes.
718 21. Refactored the implementation of \p and \P so as to be more general, to
719 allow for more different types of property in future. This has changed the
720 compiled form incompatibly. Anybody with saved compiled patterns that use
721 \p or \P will have to recompile them.
723 22. Added "Any" and "L&" to the supported property types.
725 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
726 but give a compile time error if the value is greater than 0xff.
728 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
729 accidentally not being installed or uninstalled.
731 25. The pcre.h file was built from pcre.h.in, but the only changes that were
732 made were to insert the current release number. This seemed silly, because
733 it made things harder for people building PCRE on systems that don't run
734 "configure". I have turned pcre.h into a distributed file, no longer built
735 by "configure", with the version identification directly included. There is
736 no longer a pcre.h.in file.
738 However, this change necessitated a change to the pcre-config script as
739 well. It is built from pcre-config.in, and one of the substitutions was the
740 release number. I have updated configure.ac so that ./configure now finds
741 the release number by grepping pcre.h.
743 26. Added the ability to run the tests under valgrind.
746 Version 6.4 05-Sep-05
747 ---------------------
749 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
750 "--" to be printed when multiple files were scanned, even when none of the
751 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
752 consider it to be a bug, and have restored the previous behaviour.
754 2. A couple of code tidies to get rid of compiler warnings.
756 3. The pcretest program used to cheat by referring to symbols in the library
757 whose names begin with _pcre_. These are internal symbols that are not
758 really supposed to be visible externally, and in some environments it is
759 possible to suppress them. The cheating is now confined to including
760 certain files from the library's source, which is a bit cleaner.
762 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
763 file's purpose clearer.
765 5. Reorganized pcre_ucp_findchar().
768 Version 6.3 15-Aug-05
769 ---------------------
771 1. The file libpcre.pc.in did not have general read permission in the tarball.
773 2. There were some problems when building without C++ support:
775 (a) If C++ support was not built, "make install" and "make test" still
776 tried to test it.
778 (b) There were problems when the value of CXX was explicitly set. Some
779 changes have been made to try to fix these, and ...
781 (c) --disable-cpp can now be used to explicitly disable C++ support.
783 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
784 backslash in a target when C++ was disabled. This confuses some
785 versions of "make", apparently. Using an intermediate variable solves
786 this. (Same for CPP_LOBJ.)
788 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
789 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
790 necessary on certain architectures.
792 4. Added a setting of -export-symbols-regex to the link command to remove
793 those symbols that are exported in the C sense, but actually are local
794 within the library, and not documented. Their names all begin with
795 "_pcre_". This is not a perfect job, because (a) we have to except some
796 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
797 available (and never for static libraries). I have made a note to try to
798 find a way round (a) in the future.
801 Version 6.2 01-Aug-05
802 ---------------------
804 1. There was no test for integer overflow of quantifier values. A construction
805 such as {1111111111111111} would give undefined results. What is worse, if
806 a minimum quantifier for a parenthesized subpattern overflowed and became
807 negative, the calculation of the memory size went wrong. This could have
808 led to memory overwriting.
810 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
812 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
813 operating environments where this matters.
815 4. Applied Giuseppe Maxia's patch to add additional features for controlling
816 PCRE options from within the C++ wrapper.
818 5. Named capturing subpatterns were not being correctly counted when a pattern
819 was compiled. This caused two problems: (a) If there were more than 100
820 such subpatterns, the calculation of the memory needed for the whole
821 compiled pattern went wrong, leading to an overflow error. (b) Numerical
822 back references of the form \12, where the number was greater than 9, were
823 not recognized as back references, even though there were sufficient
824 previous subpatterns.
826 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
827 versions of gcc, e.g. 2.95.4.
830 Version 6.1 21-Jun-05
831 ---------------------
833 1. There was one reference to the variable "posix" in pcretest.c that was not
834 surrounded by "#if !defined NOPOSIX".
836 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
837 the cross-check on the old pcre_info() function, for the benefit of the
838 cut-down version of PCRE that is currently imported into Exim.
840 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
841 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
842 patterns that start (?i) but I don't think that matters. The use of (?i) is
843 just an example; this all applies to the other options as well.
845 4. Since libtool seems to echo the compile commands it is issuing, the output
846 from "make" can be reduced a bit by putting "@" in front of each libtool
847 compile command.
849 5. Patch from the folks at Google for configure.in to be a bit more thorough
850 in checking for a suitable C++ installation before trying to compile the
851 C++ stuff. This should fix a reported problem when a compiler was present,
852 but no suitable headers.
854 6. The man pages all had just "PCRE" as their title. I have changed them to
855 be the relevant file name. I have also arranged that these names are
856 retained in the file doc/pcre.txt, which is a concatenation in text format
857 of all the man pages except the little individual ones for each function.
859 7. The NON-UNIX-USE file had not been updated for the different set of source
860 files that come with release 6. I also added a few comments about the C++
861 wrapper.
864 Version 6.0 07-Jun-05
865 ---------------------
867 1. Some minor internal re-organization to help with my DFA experiments.
869 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
870 didn't matter for the library itself when fully configured, but did matter
871 when compiling without UCP support, or within Exim, where the ucp files are
872 not imported.
874 3. Refactoring of the library code to split up the various functions into
875 different source modules. The addition of the new DFA matching code (see
876 below) to a single monolithic source would have made it really too
877 unwieldy, quite apart from causing all the code to be include in a
878 statically linked application, when only some functions are used. This is
879 relevant even without the DFA addition now that patterns can be compiled in
880 one application and matched in another.
882 The downside of splitting up is that there have to be some external
883 functions and data tables that are used internally in different modules of
884 the library but which are not part of the API. These have all had their
885 names changed to start with "_pcre_" so that they are unlikely to clash
886 with other external names.
888 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
889 a different (DFA) algorithm. Although it is slower than the original
890 function, it does have some advantages for certain types of matching
891 problem.
893 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
894 including restarting after a partial match.
896 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
897 defined when compiling for Windows was sent to me. I have put it into the
898 code, though I have no means of testing or verifying it.
900 7. Added the pcre_refcount() auxiliary function.
902 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
903 match before or at the first newline in the subject string. In pcretest,
904 the /f option on a pattern can be used to set this.
906 9. A repeated \w when used in UTF-8 mode with characters greater than 256
907 would behave wrongly. This has been present in PCRE since release 4.0.
909 10. A number of changes to the pcregrep command:
911 (a) Refactored how -x works; insert ^(...)$ instead of setting
912 PCRE_ANCHORED and checking the length, in preparation for adding
913 something similar for -w.
915 (b) Added the -w (match as a word) option.
917 (c) Refactored the way lines are read and buffered so as to have more
918 than one at a time available.
920 (d) Implemented a pcregrep test script.
922 (e) Added the -M (multiline match) option. This allows patterns to match
923 over several lines of the subject. The buffering ensures that at least
924 8K, or the rest of the document (whichever is the shorter) is available
925 for matching (and similarly the previous 8K for lookbehind assertions).
927 (f) Changed the --help output so that it now says
929 -w, --word-regex(p)
931 instead of two lines, one with "regex" and the other with "regexp"
932 because that confused at least one person since the short forms are the
933 same. (This required a bit of code, as the output is generated
934 automatically from a table. It wasn't just a text change.)
936 (g) -- can be used to terminate pcregrep options if the next thing isn't an
937 option but starts with a hyphen. Could be a pattern or a path name
938 starting with a hyphen, for instance.
940 (h) "-" can be given as a file name to represent stdin.
942 (i) When file names are being printed, "(standard input)" is used for
943 the standard input, for compatibility with GNU grep. Previously
944 "<stdin>" was used.
946 (j) The option --label=xxx can be used to supply a name to be used for
947 stdin when file names are being printed. There is no short form.
949 (k) Re-factored the options decoding logic because we are going to add
950 two more options that take data. Such options can now be given in four
951 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
953 (l) Added the -A, -B, and -C options for requesting that lines of context
954 around matches be printed.
956 (m) Added the -L option to print the names of files that do not contain
957 any matching lines, that is, the complement of -l.
959 (n) The return code is 2 if any file cannot be opened, but pcregrep does
960 continue to scan other files.
962 (o) The -s option was incorrectly implemented. For compatibility with other
963 greps, it now suppresses the error message for a non-existent or non-
964 accessible file (but not the return code). There is a new option called
965 -q that suppresses the output of matching lines, which was what -s was
966 previously doing.
968 (p) Added --include and --exclude options to specify files for inclusion
969 and exclusion when recursing.
971 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
972 Hopefully, it now does.
974 12. Missing cast in pcre_study().
976 13. Added an "uninstall" target to the makefile.
978 14. Replaced "extern" in the function prototypes in Makefile.in with
979 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
980 world, but is set differently for Windows.
982 15. Added a second compiling function called pcre_compile2(). The only
983 difference is that it has an extra argument, which is a pointer to an
984 integer error code. When there is a compile-time failure, this is set
985 non-zero, in addition to the error test pointer being set to point to an
986 error message. The new argument may be NULL if no error number is required
987 (but then you may as well call pcre_compile(), which is now just a
988 wrapper). This facility is provided because some applications need a
989 numeric error indication, but it has also enabled me to tidy up the way
990 compile-time errors are handled in the POSIX wrapper.
992 16. Added VPATH=.libs to the makefile; this should help when building with one
993 prefix path and installing with another. (Or so I'm told by someone who
994 knows more about this stuff than I do.)
996 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
997 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
998 match everything, including newlines. This is not POSIX-compatible, but
999 somebody wanted the feature. From pcretest it can be activated by using
1000 both the P and the s flags.
1002 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1004 19. libpcre.pc was being incorrectly installed as executable.
1006 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1007 it now also looks for '\r' so that it will work unmodified on Windows.
1009 21. Added Google's contributed C++ wrapper to the distribution.
1011 22. Added some untidy missing memory free() calls in pcretest, to keep
1012 Electric Fence happy when testing.
1016 Version 5.0 13-Sep-04
1017 ---------------------
1019 1. Internal change: literal characters are no longer packed up into items
1020 containing multiple characters in a single byte-string. Each character
1021 is now matched using a separate opcode. However, there may be more than one
1022 byte in the character in UTF-8 mode.
1024 2. The pcre_callout_block structure has two new fields: pattern_position and
1025 next_item_length. These contain the offset in the pattern to the next match
1026 item, and its length, respectively.
1028 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1029 insertion of callouts before each pattern item. Added the /C option to
1030 pcretest to make use of this.
1032 4. On the advice of a Windows user, the lines
1034 #if defined(_WIN32) || defined(WIN32)
1035 _setmode( _fileno( stdout ), 0x8000 );
1036 #endif /* defined(_WIN32) || defined(WIN32) */
1038 have been added to the source of pcretest. This apparently does useful
1039 magic in relation to line terminators.
1041 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1042 for the benefit of those environments where the "b" makes a difference.
1044 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1045 to know about it. I have put a hack into configure.in that adds in code
1046 to set GCC=yes if CC=icc. This seems to end up at a point in the
1047 generated configure script that is early enough to affect the setting of
1048 compiler options, which is what is needed, but I have no means of testing
1049 whether it really works. (The user who reported this had patched the
1050 generated configure script, which of course I cannot do.)
1052 LATER: After change 22 below (new libtool files), the configure script
1053 seems to know about icc (and also ecc). Therefore, I have commented out
1054 this hack in configure.in.
1056 7. Added support for pkg-config (2 patches were sent in).
1058 8. Negated POSIX character classes that used a combination of internal tables
1059 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1060 [[:^ascii]]. Typically, they would match almost any characters. The other
1061 POSIX classes were not broken in this way.
1063 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1064 to find the match, as PCRE was deluded into thinking that the match had to
1065 start at the start point or following a newline. The same bug applied to
1066 patterns with negative forward assertions or any backward assertions
1067 preceding ".*" at the start, unless the pattern required a fixed first
1068 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1070 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1071 starting at the last subject character, bytes beyond the end of the subject
1072 string were read.
1074 11. Renamed the variable "class" as "classbits" to make life easier for C++
1075 users. (Previously there was a macro definition, but it apparently wasn't
1076 enough.)
1078 12. Added the new field "tables" to the extra data so that tables can be passed
1079 in at exec time, or the internal tables can be re-selected. This allows
1080 a compiled regex to be saved and re-used at a later time by a different
1081 program that might have everything at different addresses.
1083 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1084 -R library as well as a -L library.
1086 14. The debugging options of pcretest (-d on the command line or D on a
1087 pattern) showed incorrect output for anything following an extended class
1088 that contained multibyte characters and which was followed by a quantifier.
1090 15. Added optional support for general category Unicode character properties
1091 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1092 support. It adds about 90K to the size of the library. The meanings of the
1093 inbuilt class escapes such as \d and \s have NOT been changed.
1095 16. Updated pcredemo.c to include calls to free() to release the memory for the
1096 compiled pattern.
1098 17. The generated file chartables.c was being created in the source directory
1099 instead of in the building directory. This caused the build to fail if the
1100 source directory was different from the building directory, and was
1101 read-only.
1103 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1104 file. No doubt somebody will tell me if they don't make sense... Also added
1105 Dan Mooney's comments about building on OpenVMS.
1107 19. Added support for partial matching via the PCRE_PARTIAL option for
1108 pcre_exec() and the \P data escape in pcretest.
1110 20. Extended pcretest with 3 new pattern features:
1112 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1113 write the compiled pattern to the file whose name is "rest-of-line".
1114 This is a straight binary dump of the data, with the saved pointer to
1115 the character tables forced to be NULL. The study data, if any, is
1116 written too. After writing, pcretest reads a new pattern.
1118 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1119 compiled pattern from the given file. There must not be any
1120 occurrences of "<" in the file name (pretty unlikely); if there are,
1121 pcretest will instead treat the initial "<" as a pattern delimiter.
1122 After reading in the pattern, pcretest goes on to read data lines as
1123 usual.
1125 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1126 and 16-bit fields in a compiled pattern, to simulate a pattern that
1127 was compiled on a host of opposite endianness.
1129 21. The pcre-exec() function can now cope with patterns that were compiled on
1130 hosts of opposite endianness, with this restriction:
1132 As for any compiled expression that is saved and used later, the tables
1133 pointer field cannot be preserved; the extra_data field in the arguments
1134 to pcre_exec() should be used to pass in a tables address if a value
1135 other than the default internal tables were used at compile time.
1137 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1138 now diagnosed as an error. Previously, most of the time, a negative number
1139 would have been treated as zero, but if in addition "ovector" was passed as
1140 NULL, a crash could occur.
1142 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1143 new versions from the libtool 1.5 distribution (the last one is a copy of
1144 a file called libtool.m4). This seems to have fixed the need to patch
1145 "configure" to support Darwin 1.3 (which I used to do). However, I still
1146 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1147 workstation).
1149 24. Changed the PCRE licence to be the more standard "BSD" licence.
1152 Version 4.5 01-Dec-03
1153 ---------------------
1155 1. There has been some re-arrangement of the code for the match() function so
1156 that it can be compiled in a version that does not call itself recursively.
1157 Instead, it keeps those local variables that need separate instances for
1158 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1159 needs to "recurse". Keeping track of where control must go is done by means
1160 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1161 hide most of the details from the main code, and operates only if
1162 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1163 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1164 operating.
1166 To make it easier for callers to provide specially tailored get/free
1167 functions for this usage, two new functions, pcre_stack_malloc, and
1168 pcre_stack_free, are used. They are always called in strict stacking order,
1169 and the size of block requested is always the same.
1171 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1172 PCRE has been compiled to use the stack or the heap for recursion. The
1173 -C option of pcretest uses this to show which version is compiled.
1175 A new data escape \S, is added to pcretest; it causes the amounts of store
1176 obtained and freed by both kinds of malloc/free at match time to be added
1177 to the output.
1179 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1180 what's available on my current Linux desktop machine.
1182 3. When matching a UTF-8 string, the test for a valid string at the start has
1183 been extended. If start_offset is not zero, PCRE now checks that it points
1184 to a byte that is the start of a UTF-8 character. If not, it returns
1185 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1186 this is necessary because there may be backward assertions in the pattern.
1187 When matching the same subject several times, it may save resources to use
1188 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1190 4. The code for checking the validity of UTF-8 strings has been tightened so
1191 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1192 containing "overlong sequences".
1194 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1195 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1196 should have been "&", but it just so happened that all the cases this let
1197 through by mistake were picked up later in the function.
1199 6. I had used a variable called "isblank" - this is a C99 function, causing
1200 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1202 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1203 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1204 so the version is shown; (c) stop "make test" echoing "./RunTest".
1206 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1208 9. The prototype for memmove() for systems that don't have it was using
1209 size_t, but the inclusion of the header that defines size_t was later. I've
1210 moved the #includes for the C headers earlier to avoid this.
1212 10. Added some adjustments to the code to make it easier to compiler on certain
1213 special systems:
1215 (a) Some "const" qualifiers were missing.
1216 (b) Added the macro EXPORT before all exported functions; by default this
1217 is defined to be empty.
1218 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1219 that it reads its output file name as an argument instead of writing
1220 to the standard output and assuming this can be redirected.
1222 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1223 class containing characters with values greater than 255, PCRE compilation
1224 went into a loop.
1226 12. A recursive reference to a subpattern that was within another subpattern
1227 that had a minimum quantifier of zero caused PCRE to crash. For example,
1228 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1229 recursion. If the recursively-called subpattern itself had a zero repeat,
1230 that was OK.
1232 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1233 buffer into which it was copied (for escape processing) was still set at
1234 1024, so long lines caused crashes.
1236 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1237 "internal error: code overflow...". This applied to any character class
1238 that was followed by a possessive quantifier.
1240 15. Modified the Makefile to add libpcre.la as a prerequisite for
1241 libpcreposix.la because I was told this is needed for a parallel build to
1242 work.
1244 16. If a pattern that contained .* following optional items at the start was
1245 studied, the wrong optimizing data was generated, leading to matching
1246 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1247 matching string must start with a or b or c. The correct conclusion for
1248 this pattern is that a match can start with any character.
1251 Version 4.4 13-Aug-03
1252 ---------------------
1254 1. In UTF-8 mode, a character class containing characters with values between
1255 127 and 255 was not handled correctly if the compiled pattern was studied.
1256 In fixing this, I have also improved the studying algorithm for such
1257 classes (slightly).
1259 2. Three internal functions had redundant arguments passed to them. Removal
1260 might give a very teeny performance improvement.
1262 3. Documentation bug: the value of the capture_top field in a callout is *one
1263 more than* the number of the hightest numbered captured substring.
1265 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1266 in incorrectly linking with a previously installed version. They now link
1267 explicitly with libpcre.la.
1269 5. configure.in no longer needs to recognize Cygwin specially.
1271 6. A problem in pcre.in for Windows platforms is fixed.
1273 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1274 pcretest, it used to include the size of the study block as part of its
1275 output. Unfortunately, the structure contains a field that has a different
1276 size on different hardware architectures. This meant that the tests that
1277 showed this size failed. As the block is currently always of a fixed size,
1278 this information isn't actually particularly useful in pcretest output, so
1279 I have just removed it.
1281 8. Three pre-processor statements accidentally did not start in column 1.
1282 Sadly, there are *still* compilers around that complain, even though
1283 standard C has not required this for well over a decade. Sigh.
1285 9. In pcretest, the code for checking callouts passed small integers in the
1286 callout_data field, which is a void * field. However, some picky compilers
1287 complained about the casts involved for this on 64-bit systems. Now
1288 pcretest passes the address of the small integer instead, which should get
1289 rid of the warnings.
1291 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1292 both compile and run time, and gives an error if an invalid UTF-8 sequence
1293 is found. There is a option for disabling this check in cases where the
1294 string is known to be correct and/or the maximum performance is wanted.
1296 11. In response to a bug report, I changed one line in Makefile.in from
1298 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1299 to
1300 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1302 to look similar to other lines, but I have no way of telling whether this
1303 is the right thing to do, as I do not use Windows. No doubt I'll get told
1304 if it's wrong...
1307 Version 4.3 21-May-03
1308 ---------------------
1310 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1311 Makefile.
1313 2. Some refactoring to improve the quality of the code:
1315 (i) The utf8_table... variables are now declared "const".
1317 (ii) The code for \cx, which used the "case flipping" table to upper case
1318 lower case letters, now just substracts 32. This is ASCII-specific,
1319 but the whole concept of \cx is ASCII-specific, so it seems
1320 reasonable.
1322 (iii) PCRE was using its character types table to recognize decimal and
1323 hexadecimal digits in the pattern. This is silly, because it handles
1324 only 0-9, a-f, and A-F, but the character types table is locale-
1325 specific, which means strange things might happen. A private
1326 table is now used for this - though it costs 256 bytes, a table is
1327 much faster than multiple explicit tests. Of course, the standard
1328 character types table is still used for matching digits in subject
1329 strings against \d.
1331 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1332 ending in _t are). So I've renamed it as ESC_tee.
1334 3. The first argument for regexec() in the POSIX wrapper should have been
1335 defined as "const".
1337 4. Changed pcretest to use malloc() for its buffers so that they can be
1338 Electric Fenced for debugging.
1340 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1341 to read one or more bytes before the start of the subject string. Often this
1342 had no effect on PCRE's behaviour, but in some circumstances it could
1343 provoke a segmentation fault.
1345 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1346 to try to read one or more bytes before the start of the subject string.
1348 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1349 UTF-8 support could misbehave in various ways if the subject string
1350 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1351 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1352 back over UTF-8 characters.)
1355 Version 4.2 14-Apr-03
1356 ---------------------
1358 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1360 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1361 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1362 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1363 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1364 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1366 Note: automatic setting of the BUILD variables is not yet working
1367 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1368 compile-time but not at link-time
1369 [LINK]: use for linking executables only
1370 make different versions for Windows and non-Windows
1371 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1372 libraries
1373 [LINK_FOR_BUILD]: new variable
1374 [OBJEXT]: use throughout
1375 [EXEEXT]: use throughout
1376 <winshared>: new target
1377 <wininstall>: new target
1378 <dftables.o>: use native compiler
1379 <dftables>: use native linker
1380 <install>: handle Windows platform correctly
1381 <clean>: ditto
1382 <check>: ditto
1383 copy DLL to top builddir before testing
1385 As part of these changes, -no-undefined was removed again. This was reported
1386 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1387 in any case.
1389 3. Some tidies to get rid of compiler warnings:
1391 . In the match_data structure, match_limit was an unsigned long int, whereas
1392 match_call_count was an int. I've made them both unsigned long ints.
1394 . In pcretest the fact that a const uschar * doesn't automatically cast to
1395 a void * provoked a warning.
1397 . Turning on some more compiler warnings threw up some "shadow" variables
1398 and a few more missing casts.
1400 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1401 option, a class that contained a single character with a value between 128
1402 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1404 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1405 option, a class that contained several characters, but with at least one
1406 whose value was between 128 and 255 caused PCRE to crash.
1409 Version 4.1 12-Mar-03
1410 ---------------------
1412 1. Compiling with gcc -pedantic found a couple of places where casts were
1413 needed, and a string in dftables.c that was longer than standard compilers are
1414 required to support.
1416 2. Compiling with Sun's compiler found a few more places where the code could
1417 be tidied up in order to avoid warnings.
1419 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1420 first of these names is deprecated in the latest Autoconf in favour of the name
1421 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1422 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1423 analogy I have changed it to CFLAGS_FOR_BUILD.
1425 4. Added -no-undefined to the linking command in the Makefile, because this is
1426 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1427 linking step for the pcreposix library.
1429 5. PCRE was failing to diagnose the case of two named groups with the same
1430 name.
1432 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1433 literal character that is needed in the subject for a match, and scans along to
1434 ensure that it is present before embarking on the full matching process. This
1435 saves time in cases of nested unlimited repeats that are never going to match.
1436 Problem: the scan can take a lot of time if the subject is very long (e.g.
1437 megabytes), thus penalizing straightforward matches. It is now done only if the
1438 amount of subject to be scanned is less than 1000 bytes.
1440 7. A lesser problem with the same optimization is that it was recording the
1441 first character of an anchored pattern as "needed", thus provoking a search
1442 right along the subject, even when the first match of the pattern was going to
1443 fail. The "needed" character is now not set for anchored patterns, unless it
1444 follows something in the pattern that is of non-fixed length. Thus, it still
1445 fulfils its original purpose of finding quick non-matches in cases of nested
1446 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1449 Version 4.0 17-Feb-03
1450 ---------------------
1452 1. If a comment in an extended regex that started immediately after a meta-item
1453 extended to the end of string, PCRE compiled incorrect data. This could lead to
1454 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1456 2. Moved to autoconf 2.53 and libtool 1.4.2.
1458 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1459 the special perltest8 script is no longer needed - all the tests can be run
1460 from a single perltest script.
1462 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1463 by \s. It has now been removed in PCRE. This means it isn't recognized as
1464 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1465 class [:space:] *does* include VT, thereby creating a mess.
1467 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1468 space and tab.
1470 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1471 its new features into the main test script, reducing the number of scripts.
1473 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1474 were backward compatible, and made the (?i) apply to the whole pattern, as if
1475 /i were given. Now it behaves more logically, and applies the option setting
1476 only to what follows. PCRE has been changed to follow suit. However, if it
1477 finds options settings right at the start of the pattern, it extracts them into
1478 the global options, as before. Thus, they show up in the info data.
1480 8. Added support for the \Q...\E escape sequence. Characters in between are
1481 treated as literals. This is slightly different from Perl in that $ and @ are
1482 also handled as literals inside the quotes. In Perl, they will cause variable
1483 interpolation. Note the following examples:
1485 Pattern PCRE matches Perl matches
1487 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1488 \Qabc\$xyz\E abc\$xyz abc\$xyz
1489 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1491 For compatibility with Perl, \Q...\E sequences are recognized inside character
1492 classes as well as outside them.
1494 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1495 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1496 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1497 signed/unsigned warnings.
1499 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1500 option for pcretest, so I've replaced it by a simple function that does just
1501 that job.
1503 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1504 "pcregrep -".
1506 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1507 Java package. This provides some syntactic sugar for simple cases of what my
1508 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1509 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1510 item, you can use this simplified notation. Note that only makes sense with
1511 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1512 greediness, whatever the setting of the PCRE_UNGREEDY option.
1514 13. A change of greediness default within a pattern was not taking effect at
1515 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1516 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1517 was abstracted outside.
1519 14. PCRE now supports the \G assertion. It is true when the current matching
1520 position is at the start point of the match. This differs from \A when the
1521 starting offset is non-zero. Used with the /g option of pcretest (or similar
1522 code), it works in the same way as it does for Perl's /g option. If all
1523 alternatives of a regex begin with \G, the expression is anchored to the start
1524 match position, and the "anchored" flag is set in the compiled expression.
1526 15. Some bugs concerning the handling of certain option changes within patterns
1527 have been fixed. These applied to options other than (?ims). For example,
1528 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1529 been the other way round. Some of this was related to change 7 above.
1531 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1532 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1533 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1534 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1536 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1537 mode. Unlike ".", it always matches newline, whatever the setting of
1538 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1539 assertions. Perl allows it, but it doesn't (in general) work because it can't
1540 calculate the length of the lookbehind. At least, that's the case for Perl
1541 5.8.0 - I've been told they are going to document that it doesn't work in
1542 future.
1544 18. Added an error diagnosis for escapes that PCRE does not support: these are
1545 \L, \l, \N, \P, \p, \U, \u, and \X.
1547 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1548 reading past the end of the pattern in cases such as /[abcd/.
1550 20. PCRE was getting more memory than necessary for patterns with classes that
1551 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1553 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1554 compiling PCRE for use with Virtual Pascal.
1556 22. Small fix to the Makefile to make it work properly if the build is done
1557 outside the source tree.
1559 23. Added a new extension: a condition to go with recursion. If a conditional
1560 subpattern starts with (?(R) the "true" branch is used if recursion has
1561 happened, whereas the "false" branch is used only at the top level.
1563 24. When there was a very long string of literal characters (over 255 bytes
1564 without UTF support, over 250 bytes with UTF support), the computation of how
1565 much memory was required could be incorrect, leading to segfaults or other
1566 strange effects.
1568 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1569 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1570 there was a subsequent back reference to those brackets. This meant that, for
1571 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1572 possible to check for precisely this case. All we can do is abandon the
1573 optimization if .* occurs inside capturing brackets when there are any back
1574 references whatsoever. (See below for a better fix that came later.)
1576 26. The handling of the optimization for finding the first character of a
1577 non-anchored pattern, and for finding a character that is required later in the
1578 match were failing in some cases. This didn't break the matching; it just
1579 failed to optimize when it could. The way this is done has been re-implemented.
1581 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1583 28. Added a new feature that provides some of the functionality that Perl
1584 provides with (?{...}). The facility is termed a "callout". The way it is done
1585 in PCRE is for the caller to provide an optional function, by setting
1586 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1587 global variable. By default it is unset, which disables all calling out. To get
1588 the function called, the regex must include (?C) at appropriate points. This
1589 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1590 This provides a means of identifying different callout points. When PCRE
1591 reaches such a point in the regex, if pcre_callout has been set, the external
1592 function is called. It is provided with data in a structure called
1593 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1594 matching continues; if it returns a non-zero value, the match at the current
1595 point fails. However, backtracking will occur if possible. [This was changed
1596 later and other features added - see item 49 below.]
1598 29. pcretest is upgraded to test the callout functionality. It provides a
1599 callout function that displays information. By default, it shows the start of
1600 the match and the current position in the text. There are some new data escapes
1601 to vary what happens:
1603 \C+ in addition, show current contents of captured substrings
1604 \C- do not supply a callout function
1605 \C!n return 1 when callout number n is reached
1606 \C!n!m return 1 when callout number n is reached for the mth time
1608 30. If pcregrep was called with the -l option and just a single file name, it
1609 output "<stdin>" if a match was found, instead of the file name.
1611 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1612 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1613 pcre_exec(). This saves a malloc/free per call. The default value of
1614 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1615 when configuring.
1617 32. The default maximum size of a compiled pattern is 64K. There have been a
1618 few cases of people hitting this limit. The code now uses macros to handle the
1619 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1620 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1621 configuring. Tests 2 and 5 work only with 2-byte links because they output
1622 debugging information about compiled patterns.
1624 33. Internal code re-arrangements:
1626 (a) Moved the debugging function for printing out a compiled regex into
1627 its own source file (printint.c) and used #include to pull it into
1628 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1629 separate copies.
1631 (b) Defined the list of op-code names for debugging as a macro in
1632 internal.h so that it is next to the definition of the opcodes.
1634 (c) Defined a table of op-code lengths for simpler skipping along compiled
1635 code. This is again a macro in internal.h so that it is next to the
1636 definition of the opcodes.
1638 34. Added support for recursive calls to individual subpatterns, along the
1639 lines of Robin Houston's patch (but implemented somewhat differently).
1641 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1642 allow it to read and process whole directories in Win32. This code was
1643 contributed by Lionel Fourquaux; it has not been tested by me.
1645 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1646 used to name a group. Names consist of alphanumerics and underscores, and must
1647 be unique. Back references use the syntax (?P=name) and recursive calls use
1648 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1649 numbers. The function pcre_fullinfo() can be used after compilation to extract
1650 a name/number map. There are three relevant calls:
1652 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1653 PCRE_INFO_NAMECOUNT yields the number of entries
1654 PCRE_INFO_NAMETABLE yields a pointer to the map.
1656 The map is a vector of fixed-size entries. The size of each entry depends on
1657 the length of the longest name used. The first two bytes of each entry are the
1658 group number, most significant byte first. There follows the corresponding
1659 name, zero terminated. The names are in alphabetical order.
1661 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1662 case instead of 255. Making it the same both with and without UTF-8 support
1663 means that the same test output works with both.
1665 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1666 calling malloc() with a zero argument.
1668 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1669 optimization. I've improved things by keeping a bitmap of backreferences with
1670 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1671 fact referenced, the optimization can be applied. It is unlikely that a
1672 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1673 the match to follow \n) will appear inside brackets with a number greater than
1674 31, but if it does, any back reference > 31 suppresses the optimization.
1676 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1677 of disabling numbered capturing parentheses. Any opening parenthesis that is
1678 not followed by ? behaves as if it were followed by ?: but named parentheses
1679 can still be used for capturing (and they will acquire numbers in the usual
1680 way).
1682 41. Redesigned the return codes from the match() function into yes/no/error so
1683 that errors can be passed back from deep inside the nested calls. A malloc
1684 failure while inside a recursive subpattern call now causes the
1685 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1687 42. It is now possible to set a limit on the number of times the match()
1688 function is called in a call to pcre_exec(). This facility makes it possible to
1689 limit the amount of recursion and backtracking, though not in a directly
1690 obvious way, because the match() function is used in a number of different
1691 circumstances. The count starts from zero for each position in the subject
1692 string (for non-anchored patterns). The default limit is, for compatibility, a
1693 large number, namely 10 000 000. You can change this in two ways:
1695 (a) When configuring PCRE before making, you can use --with-match-limit=n
1696 to set a default value for the compiled library.
1698 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1699 a different value is set. See 45 below.
1701 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1703 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1704 of things that can be changed at compile time. The first argument specifies
1705 what is wanted and the second points to where the information is to be placed.
1706 The current list of available information is:
1710 The output is an integer that is set to one if UTF-8 support is available;
1711 otherwise it is set to zero.
1715 The output is an integer that it set to the value of the code that is used for
1716 newline. It is either LF (10) or CR (13).
1720 The output is an integer that contains the number of bytes used for internal
1721 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1725 The output is an integer that contains the threshold above which the POSIX
1726 interface uses malloc() for output vectors. See item 31 above.
1730 The output is an unsigned integer that contains the default limit of the number
1731 of match() calls in a pcre_exec() execution. See 42 above.
1733 44. pcretest has been upgraded by the addition of the -C option. This causes it
1734 to extract all the available output from the new pcre_config() function, and to
1735 output it. The program then exits immediately.
1737 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1738 order to support additional features. One way would have been to define
1739 pcre_exec2() (for example) with extra arguments, but this would not have been
1740 extensible, and would also have required all calls to the original function to
1741 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1742 is used for passing in "extra" data from pcre_study().
1744 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1745 contains the following fields:
1747 flags a bitmap indicating which of the following fields are set
1748 study_data opaque data from pcre_study()
1749 match_limit a way of specifying a limit on match() calls for a specific
1750 call to pcre_exec()
1751 callout_data data for callouts (see 49 below)
1753 The flag bits are also defined in pcre.h, and are
1759 The pcre_study() function now returns one of these new pcre_extra blocks, with
1760 the actual study data pointed to by the study_data field, and the
1761 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1762 before. That is, this change is entirely upwards-compatible and requires no
1763 change to existing code.
1765 If you want to pass in additional data to pcre_exec(), you can either place it
1766 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1767 block.
1769 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1770 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1771 times with different match limits, until it finds the minimum value needed for
1772 pcre_exec() to complete. The value is then output. This can be instructive; for
1773 most simple matches the number is quite small, but for pathological cases it
1774 gets very large very quickly.
1776 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1777 returns the size of the data block pointed to by the study_data field in a
1778 pcre_extra block, that is, the value that was passed as the argument to
1779 pcre_malloc() when PCRE was getting memory in which to place the information
1780 created by pcre_study(). The fourth argument should point to a size_t variable.
1781 pcretest has been extended so that this information is shown after a successful
1782 pcre_study() call when information about the compiled regex is being displayed.
1784 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1785 because what follows is always an absolute path. (Later: it turns out that this
1786 is more than cosmetic for MinGW, because it doesn't like empty path
1787 components.)
1789 49. Some changes have been made to the callout feature (see 28 above):
1791 (i) A callout function now has three choices for what it returns:
1793 0 => success, carry on matching
1794 > 0 => failure at this point, but backtrack if possible
1795 < 0 => serious error, return this value from pcre_exec()
1797 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1798 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1799 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1800 use by callout functions. It will never be used by PCRE itself.
1802 (ii) The pcre_extra structure (see 45 above) has a void * field called
1803 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1804 pcre_callout_block structure has a field of the same name. The contents of
1805 the field passed in the pcre_extra structure are passed to the callout
1806 function in the corresponding field in the callout block. This makes it
1807 easier to use the same callout-containing regex from multiple threads. For
1808 testing, the pcretest program has a new data escape
1810 \C*n pass the number n (may be negative) as callout_data
1812 If the callout function in pcretest receives a non-zero value as
1813 callout_data, it returns that value.
1815 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1816 there were some redundant $(CFLAGS) in commands that are now specified as
1817 $(LINK), which already includes $(CFLAGS).
1819 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1820 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1821 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1822 one-byte characters throughout. Note that case-insensitive matching applies
1823 only to characters whose values are less than 256. PCRE doesn't support the
1824 notion of cases for higher-valued characters.
1826 (i) A character class whose characters are all within 0-255 is handled as
1827 a bit map, and the map is inverted for negative classes. Previously, a
1828 character > 255 always failed to match such a class; however it should
1829 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1831 (ii) A negated character class with a single character < 255 is coded as
1832 "not this character" (OP_NOT). This wasn't working properly when the test
1833 character was multibyte, either singly or repeated.
1835 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1836 mode, for example: \x{100}{2,3}.
1838 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1839 singly or repeated) now correctly test multibyte characters. However,
1840 PCRE doesn't recognize any characters with values greater than 255 as
1841 digits, spaces, or word characters. Such characters always match \D, \S,
1842 and \W, and never match \d, \s, or \w.
1844 (v) Classes may now contain characters and character ranges with values
1845 greater than 255. For example: [ab\x{100}-\x{400}].
1847 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1848 PCRE in UTF-8 mode.
1850 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1851 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1852 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1853 value.)
1855 53. The single man page has become too large. I have therefore split it up into
1856 a number of separate man pages. These also give rise to individual HTML pages;
1857 these are now put in a separate directory, and there is an index.html page that
1858 lists them all. Some hyperlinking between the pages has been installed.
1860 54. Added convenience functions for handling named capturing parentheses.
1862 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1863 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1864 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1865 are faulted.
1867 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1868 calling configure. These values are used when compiling the dftables.c program
1869 which is run to generate the source of the default character tables. They
1870 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1871 you will need to set these values.
1873 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1876 Version 3.9 02-Jan-02
1877 ---------------------
1879 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1881 2. If --disable-static was given, the building process failed when trying to
1882 build pcretest and pcregrep. (For some reason it was using libtool to compile
1883 them, which is not right, as they aren't part of the library.)
1886 Version 3.8 18-Dec-01
1887 ---------------------
1889 1. The experimental UTF-8 code was completely screwed up. It was packing the
1890 bytes in the wrong order. How dumb can you get?
1893 Version 3.7 29-Oct-01
1894 ---------------------
1896 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1897 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1898 this didn't happen under Solaris 8, where I normally test things.
1900 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1901 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1902 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1903 this fix an hour or so after the initial 3.7 release.)
1906 Version 3.6 23-Oct-01
1907 ---------------------
1909 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1910 offsets passed as NULL with zero offset count.
1912 2. The config.guess and config.sub files had not been updated when I moved to
1913 the latest autoconf.
1916 Version 3.5 15-Aug-01
1917 ---------------------
1919 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1920 had been forgotten.
1922 2. By using declared but undefined structures, we can avoid using "void"
1923 definitions in pcre.h while keeping the internal definitions of the structures
1924 private.
1926 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
1927 user point of view, this means that both static and shared libraries are built
1928 by default, but this can be individually controlled. More of the work of
1929 handling this static/shared cases is now inside libtool instead of PCRE's make
1930 file.
1932 4. The pcretest utility is now installed along with pcregrep because it is
1933 useful for users (to test regexs) and by doing this, it automatically gets
1934 relinked by libtool. The documentation has been turned into a man page, so
1935 there are now .1, .txt, and .html versions in /doc.
1937 5. Upgrades to pcregrep:
1938 (i) Added long-form option names like gnu grep.
1939 (ii) Added --help to list all options with an explanatory phrase.
1940 (iii) Added -r, --recursive to recurse into sub-directories.
1941 (iv) Added -f, --file to read patterns from a file.
1943 6. pcre_exec() was referring to its "code" argument before testing that
1944 argument for NULL (and giving an error if it was NULL).
1946 7. Upgraded Makefile.in to allow for compiling in a different directory from
1947 the source directory.
1949 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
1950 options bits, the pointer it was passed was to an int instead of to an unsigned
1951 long int. This mattered only on 64-bit systems.
1953 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
1954 generated) instead of pcre.in, which it its source. Also made the same change
1955 in several of the .c files.
1957 10. A new release of gcc defines printf() as a macro, which broke pcretest
1958 because it had an ifdef in the middle of a string argument for printf(). Fixed
1959 by using separate calls to printf().
1961 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
1962 script, to force use of CR or LF instead of \n in the source. On non-Unix
1963 systems, the value can be set in config.h.
1965 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
1966 absolute limit. Changed the text of the error message to make this clear, and
1967 likewise updated the man page.
1969 13. The limit of 99 on the number of capturing subpatterns has been removed.
1970 The new limit is 65535, which I hope will not be a "real" limit.
1973 Version 3.4 22-Aug-00
1974 ---------------------
1976 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
1978 2. Diagnose condition (?(0) as an error instead of crashing on matching.
1981 Version 3.3 01-Aug-00
1982 ---------------------
1984 1. If an octal character was given, but the value was greater than \377, it
1985 was not getting masked to the least significant bits, as documented. This could
1986 lead to crashes in some systems.
1988 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
1989 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
1991 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
1992 These just pass their arguments on to (pcre_free)(), but they are provided
1993 because some uses of PCRE bind it to non-C systems that can call its functions,
1994 but cannot call free() or pcre_free() directly.
1996 4. Add "make test" as a synonym for "make check". Corrected some comments in
1997 the Makefile.
1999 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2000 Makefile.
2002 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2003 command called pgrep for grepping around the active processes.
2005 7. Added the beginnings of support for UTF-8 character strings.
2007 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2008 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2009 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2010 out for the ar command.)
2013 Version 3.2 12-May-00
2014 ---------------------
2016 This is purely a bug fixing release.
2018 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2019 of ZA. This was just one example of several cases that could provoke this bug,
2020 which was introduced by change 9 of version 2.00. The code for breaking
2021 infinite loops after an iteration that matches an empty string was't working
2022 correctly.
2024 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2025 when matched against abbab (for example). After matching an empty string, it
2026 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2027 caused it to match further down the string than it should.
2029 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2030 was there because it doesn't seem to be needed, and it causes trouble on some
2031 systems, as it is not a Standard C header. It has been removed.
2033 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2034 were reported on the Macintosh. The changes were from
2036 while ((c = *(++ptr)) != 0 && c != '\n');
2037 to
2038 while ((c = *(++ptr)) != 0 && c != '\n') ;
2040 Totally extraordinary, but if that's what it takes...
2042 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2043 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2044 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2045 assumes the way PCRE uses memmove() (always moving upwards).
2047 6. PCRE is being used in one environment where strchr() is not available. There
2048 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2049 faster code anyway.
2052 Version 3.1 09-Feb-00
2053 ---------------------
2055 The only change in this release is the fixing of some bugs in Makefile.in for
2056 the "install" target:
2058 (1) It was failing to install pcreposix.h.
2060 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2063 Version 3.0 01-Feb-00
2064 ---------------------
2066 1. Add support for the /+ modifier to perltest (to output $` like it does in
2067 pcretest).
2069 2. Add support for the /g modifier to perltest.
2071 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2072 matches null strings.
2074 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2075 pattern. Perl treats empty patterns specially - it reuses the most recent
2076 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2077 effect.
2079 5. The POSIX interface was broken in that it was just handing over the POSIX
2080 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2081 required a bigger vector, with some working space on the end. This means that
2082 the POSIX wrapper now has to get and free some memory, and copy the results.
2084 6. Added some simple autoconf support, placing the test data and the
2085 documentation in separate directories, re-organizing some of the
2086 information files, and making it build pcre-config (a GNU standard). Also added
2087 libtool support for building PCRE as a shared library, which is now the
2088 default.
2090 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2091 09 are not valid octal constants. Single digits will be used for minor values
2092 less than 10.
2094 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2095 existing programs that set these in the POSIX interface can use PCRE without
2096 modification.
2098 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2099 return all that pcre_info() returns, plus additional data. The pcre_info()
2100 function is retained for compatibility, but is considered to be obsolete.
2102 10. Added experimental recursion feature (?R) to handle one common case that
2103 Perl 5.6 will be able to do with (?p{...}).
2105 11. Added support for POSIX character classes like [:alpha:], which Perl is
2106 adopting.
2109 Version 2.08 31-Aug-99
2110 ----------------------
2112 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2113 trying to match at the startoffset position, but instead was moving forward to
2114 the next newline as if a previous match had failed.
2116 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2117 and could get into a loop if a null string was matched other than at the start
2118 of the subject.
2120 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2121 be distinguished at compile time, and for completeness also added PCRE_DATE.
2123 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2124 in GnuWin32 environments.
2127 Version 2.07 29-Jul-99
2128 ----------------------
2130 1. The documentation is now supplied in plain text form and HTML as well as in
2131 the form of man page sources.
2133 2. C++ compilers don't like assigning (void *) values to other pointer types.
2134 In particular this affects malloc(). Although there is no problem in Standard
2135 C, I've put in casts to keep C++ compilers happy.
2137 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2138 should be (const char *).
2140 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2141 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2142 However, I haven't made this a standard facility. The documentation doesn't
2143 mention it, and the Makefile doesn't support it.
2145 5. The Makefile now contains an "install" target, with editable destinations at
2146 the top of the file. The pcretest program is not installed.
2148 6. pgrep -V now gives the PCRE version number and date.
2150 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2151 causing the entire string to be ignored, instead of just the last character.
2153 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2154 non-matching string, it can take a very, very long time, even for strings of
2155 quite modest length, because of the nested recursion. PCRE now does better in
2156 some of these cases. It does this by remembering the last required literal
2157 character in the pattern, and pre-searching the subject to ensure it is present
2158 before running the real match. In other words, it applies a heuristic to detect
2159 some types of certain failure quickly, and in the above example, if presented
2160 with a string that has no trailing " it gives "no match" very quickly.
2162 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2163 other alternatives are tried instead.
2166 Version 2.06 09-Jun-99
2167 ----------------------
2169 1. Change pcretest's output for amount of store used to show just the code
2170 space, because the remainder (the data block) varies in size between 32-bit and
2171 64-bit systems.
2173 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2174 start matching at. This allows lookbehinds to work when searching for multiple
2175 occurrences in a string.
2177 3. Added additional options to pcretest for testing multiple occurrences:
2179 /+ outputs the rest of the string that follows a match
2180 /g loops for multiple occurrences, using the new startoffset argument
2181 /G loops for multiple occurrences by passing an incremented pointer
2183 4. PCRE wasn't doing the "first character" optimization for patterns starting
2184 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2185 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2186 the letter 'x'. On long subject strings, this gives a significant speed-up.
2189 Version 2.05 21-Apr-99
2190 ----------------------
2192 1. Changed the type of magic_number from int to long int so that it works
2193 properly on 16-bit systems.
2195 2. Fixed a bug which caused patterns starting with .* not to work correctly
2196 when the subject string contained newline characters. PCRE was assuming
2197 anchoring for such patterns in all cases, which is not correct because .* will
2198 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2199 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2200 must be retried after every newline in the subject.
2203 Version 2.04 18-Feb-99
2204 ----------------------
2206 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2207 computation of the store needed to hold the pattern was incorrect (too large).
2208 If such patterns were nested a few deep, this could multiply and become a real
2209 problem.
2211 2. Added /M option to pcretest to show the memory requirement of a specific
2212 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2214 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2215 compiled in such a way that the backtracking after subsequent failure was
2216 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2217 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2220 Version 2.03 02-Feb-99
2221 ----------------------
2223 1. Fixed typo and small mistake in man page.
2225 2. Added 4th condition (GPL supersedes if conflict) and created separate
2226 LICENCE file containing the conditions.
2228 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2229 Perl, that is the internal \ allows the delimiter to be included in the
2230 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2231 the final delimiter, add \ to the end of the pattern (to test the error).
2233 4. Added the convenience functions for extracting substrings after a successful
2234 match. Updated pcretest to make it able to test these functions.
2237 Version 2.02 14-Jan-99
2238 ----------------------
2240 1. Initialized the working variables associated with each extraction so that
2241 their saving and restoring doesn't refer to uninitialized store.
2243 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2244 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2245 fix the problem.
2247 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2248 calls, and wasn't printing the correct value for compiling calls. Increased the
2249 default value of LOOPREPEAT, and the number of significant figures in the
2250 times.
2252 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2254 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2255 a building problem on Windows NT with a FAT file system.
2258 Version 2.01 21-Oct-98
2259 ----------------------
2261 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2262 to character tables built by pcre_maketables() in the current locale. If NULL
2263 is passed, the default tables are used.
2266 Version 2.00 24-Sep-98
2267 ----------------------
2269 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2270 it any more.
2272 2. Allow quantification of (?>) groups, and make it work correctly.
2274 3. The first character computation wasn't working for (?>) groups.
2276 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2277 end of the subject) and add 5.005's \z, which really does match only at the
2278 very end of the subject.
2280 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2282 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2283 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2284 localized options. All options to pcre_study() were also removed.
2286 7. Add other new features from 5.005:
2288 $(?<= positive lookbehind
2289 $(?<! negative lookbehind
2290 (?imsx-imsx) added the unsetting capability
2291 such a setting is global if at outer level; local otherwise
2292 (?imsx-imsx:) non-capturing groups with option setting
2293 (?(cond)re|re) conditional pattern matching
2295 A backreference to itself in a repeated group matches the previous
2296 captured string.
2298 8. General tidying up of studying (both automatic and via "study")
2299 consequential on the addition of new assertions.
2301 9. As in 5.005, unlimited repeated groups that could match an empty substring
2302 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2303 runtime if any iteration does actually match an empty substring.
2305 10. Include the RunTest script in the distribution.
2307 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2308 discrepancies, some of which were old and were also with respect to 5.004. They
2309 have now been fixed.
2312 Version 1.09 28-Apr-98
2313 ----------------------
2315 1. A negated single character class followed by a quantifier with a minimum
2316 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2317 program crashes, or just wrong answers. This did not apply to negated classes
2318 containing more than one character, or to minima other than one.
2321 Version 1.08 27-Mar-98
2322 ----------------------
2324 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2326 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2327 latter must appear before anything that relies on it in the pattern.
2330 Version 1.07 16-Feb-98
2331 ----------------------
2333 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2334 repeat of a potentially empty string).
2337 Version 1.06 23-Jan-98
2338 ----------------------
2340 1. Added Markus Oberhumer's little patches for C++.
2342 2. Literal strings longer than 255 characters were broken.
2345 Version 1.05 23-Dec-97
2346 ----------------------
2348 1. Negated character classes containing more than one character were failing if
2349 PCRE_CASELESS was set at run time.
2352 Version 1.04 19-Dec-97
2353 ----------------------
2355 1. Corrected the man page, where some "const" qualifiers had been omitted.
2357 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2358 input syntax.
2360 3. Fixed memory leak which occurred when a regex with back references was
2361 matched with an offsets vector that wasn't big enough. The temporary memory
2362 that is used in this case wasn't being freed if the match failed.
2364 4. Tidied pcretest to ensure it frees memory that it gets.
2366 5. Temporary memory was being obtained in the case where the passed offsets
2367 vector was exactly big enough.
2369 6. Corrected definition of offsetof() from change 5 below.
2371 7. I had screwed up change 6 below and broken the rules for the use of
2372 setjmp(). Now fixed.
2375 Version 1.03 18-Dec-97
2376 ----------------------
2378 1. A erroneous regex with a missing opening parenthesis was correctly
2379 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2380 on some systems.
2382 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2383 it was reported that one broken compiler failed on the former because "code" is
2384 also an independent variable.
2386 3. The erroneous regex a[]b caused an array overrun reference.
2388 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2389 fail on data ending with that character. (It was going on too far, and checking
2390 the next character, typically a binary zero.) This was specific to the
2391 optimized code for single-character negative classes.
2393 5. Added a contributed patch from the TIN world which does the following:
2395 + Add an undef for memmove, in case the the system defines a macro for it.
2397 + Add a definition of offsetof(), in case there isn't one. (I don't know
2398 the reason behind this - offsetof() is part of the ANSI standard - but
2399 it does no harm).
2401 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2402 most of the places where whitespace preceded '#'. I have given up and
2403 allowed the remaining 2 cases to be at the margin.
2405 + Rename some variables in pcre to eliminate shadowing. This seems very
2406 pedantic, but does no harm, of course.
2408 6. Moved the call to setjmp() into its own function, to get rid of warnings
2409 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2411 7. Constructs such as \d{8,} were compiling into the equivalent of
2412 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2413 outcome, but in this particular case used more store than had been allocated,
2414 which caused the bug to be discovered because it threw up an internal error.
2416 8. The debugging code in both pcre and pcretest for outputting the compiled
2417 form of a regex was going wrong in the case of back references followed by
2418 curly-bracketed repeats.
2421 Version 1.02 12-Dec-97
2422 ----------------------
2424 1. Typos in pcre.3 and comments in the source fixed.
2426 2. Applied a contributed patch to get rid of places where it used to remove
2427 'const' from variables, and fixed some signed/unsigned and uninitialized
2428 variable warnings.
2430 3. Added the "runtest" target to Makefile.
2432 4. Set default compiler flag to -O2 rather than just -O.
2435 Version 1.01 19-Nov-97
2436 ----------------------
2438 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2439 like /([ab]*)*/, that is, for classes with more than one character in them.
2441 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2442 as /((?>a*))*/ (a PCRE_EXTRA facility).
2445 Version 1.00 18-Nov-97
2446 ----------------------
2448 1. Added compile-time macros to support systems such as SunOS4 which don't have
2449 memmove() or strerror() but have other things that can be used instead.
2451 2. Arranged that "make clean" removes the executables.
2454 Version 0.99 27-Oct-97
2455 ----------------------
2457 1. Fixed bug in code for optimizing classes with only one character. It was
2458 initializing a 32-byte map regardless, which could cause it to run off the end
2459 of the memory it had got.
2461 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2464 Version 0.98 22-Oct-97
2465 ----------------------
2467 1. Fixed bug in code for handling temporary memory usage when there are more
2468 back references than supplied space in the ovector. This could cause segfaults.
2471 Version 0.97 21-Oct-97
2472 ----------------------
2474 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2476 2. Optimized negated single characters not to use a bit map.
2478 3. Brought error texts together as macro definitions; clarified some of them;
2479 fixed one that was wrong - it said "range out of order" when it meant "invalid
2480 escape sequence".
2482 4. Changed some char * arguments to const char *.
2484 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2486 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2487 pcretest.
2490 Version 0.96 16-Oct-97
2491 ----------------------
2493 1. Added a simple "pgrep" utility to the distribution.
2495 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2496 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2497 where "ddd" means "one or more decimal digits".
2499 3. Fixed serious bug. If a pattern had a back reference, but the call to
2500 pcre_exec() didn't supply a large enough ovector to record the related
2501 identifying subpattern, the match always failed. PCRE now remembers the number
2502 of the largest back reference, and gets some temporary memory in which to save
2503 the offsets during matching if necessary, in order to ensure that
2504 backreferences always work.
2506 4. Increased the compatibility with Perl in a number of ways:
2508 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2509 to request this handling. The option can be set at compile or exec time.
2511 (b) $ matches before a terminating newline by default; an option
2512 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2513 mode). The option can be set at compile or exec time.
2515 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2516 the same as Perl's. If the decimal number it represents is less than 10
2517 or there aren't that many previous left capturing parentheses, an octal
2518 escape is read. Inside a character class, it's always an octal escape,
2519 even if it is a single digit.
2521 (d) An escaped but undefined alphabetic character is taken as a literal,
2522 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2523 escapes.
2525 (e) {0} is now permitted. (The previous item is removed from the compiled
2526 pattern).
2528 5. Changed all the names of code files so that the basic parts are no longer
2529 than 10 characters, and abolished the teeny "globals.c" file.
2531 6. Changed the handling of character classes; they are now done with a 32-byte
2532 bit map always.
2534 7. Added the -d and /D options to pcretest to make it possible to look at the
2535 internals of compilation without having to recompile pcre.
2538 Version 0.95 23-Sep-97
2539 ----------------------
2541 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2542 \x20 at the start of a run of normal characters. These were being treated as
2543 real characters, instead of the source characters being re-checked.
2546 Version 0.94 18-Sep-97
2547 ----------------------
2549 1. The functions are now thread-safe, with the caveat that the global variables
2550 containing pointers to malloc() and free() or alternative functions are the
2551 same for all threads.
2553 2. Get pcre_study() to generate a bitmap of initial characters for non-
2554 anchored patterns when this is possible, and use it if passed to pcre_exec().
2557 Version 0.93 15-Sep-97
2558 ----------------------
2560 1. /(b)|(:+)/ was computing an incorrect first character.
2562 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2563 but not actually doing anything yet.
2565 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2566 as Perl does (e.g. [-az] or [az-]).
2568 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2569 all possible positions.
2571 5. Split up into different modules to avoid including unneeded functions in a
2572 compiled binary. However, compile and exec are still in one module. The "study"
2573 function is split off.
2575 6. The character tables are now in a separate module whose source is generated
2576 by an auxiliary program - but can then be edited by hand if required. There are
2577 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2578 toupper() in the code.
2580 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2581 make them global. Abolish the function for setting them, as the caller can now
2582 set them directly.
2585 Version 0.92 11-Sep-97
2586 ----------------------
2588 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2589 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2591 2. Caseless matching was not working in character classes if the characters in
2592 the pattern were in upper case.
2594 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2596 4. Make PCRE_ANCHORED public and accept as a compile option.
2598 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2599 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2600 pass them.
2602 6. Give an error if bad option bits passed at compile or run time.
2604 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2605 pcretest to cause it to pass that flag.
2607 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2608 options, and the first character, if set.
2610 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2613 Version 0.91 10-Sep-97
2614 ----------------------
2616 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2617 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2619 2. PCRE was looping on encountering an indefinitely repeated back reference to
2620 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2621 Perl does - treats the match as successful.
2623 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5