ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 349 - (show annotations)
Wed Jul 2 18:42:11 2008 UTC (12 years, 7 months ago) by ph10
File size: 156588 byte(s)
Preliminary patch for upgraded Unicode Property support.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.0 02 Jul-08
5 ---------------------
7 1. Replaced UCP searching code with optimized version as implemented for Ad
8 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
9 stage table and inline lookup instead of a function, giving speed ups of 2
10 to 5 times on some simple patterns that I tested. Permission was given to
11 distribute the MultiStage2.py script that generates the tables (it's not in
12 the tarball, but is in the Subversion repository).
15 Version 7.7 07-May-08
16 ---------------------
18 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
19 a string to a long long, pretend we don't even have a long long." This is
20 done by checking for the strtoq, strtoll, and _strtoi64 functions.
22 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
23 pre-7.6 versions, which defined a global no_arg variable instead of putting
24 it in the RE class. (See also #8 below.)
26 3. Remove a line of dead code, identified by coverity and reported by Nuno
27 Lopes.
29 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
31 (1) The include/exclude patterns were being applied to the whole pathnames
32 of files, instead of just to the final components.
34 (2) If there was more than one level of directory, the subdirectories were
35 skipped unless they satisfied the include/exclude conditions. This is
36 inconsistent with GNU grep (and could even be seen as contrary to the
37 pcregrep specification - which I improved to make it absolutely clear).
38 The action now is always to scan all levels of directory, and just
39 apply the include/exclude patterns to regular files.
41 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
42 --exclude_dir in the tests to avoid scanning .svn directories.
44 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
45 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
46 doesn't support NULs in patterns.
48 7. Added some missing "const"s to declarations of static tables in
49 pcre_compile.c and pcre_dfa_exec.c.
51 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
52 caused by fix #2 above. (Subsequently also a second patch to fix the
53 first patch. And a third patch - this was a messy problem.)
55 9. Applied Craig's patch to remove the use of push_back().
57 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
58 matching function regexec().
60 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
61 which, however, unlike Perl's \g{...}, are subroutine calls, not back
62 references. PCRE supports relative numbers with this syntax (I don't think
63 Oniguruma does).
65 12. Previously, a group with a zero repeat such as (...){0} was completely
66 omitted from the compiled regex. However, this means that if the group
67 was called as a subroutine from elsewhere in the pattern, things went wrong
68 (an internal error was given). Such groups are now left in the compiled
69 pattern, with a new opcode that causes them to be skipped at execution
70 time.
72 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
73 to the way PCRE behaves:
75 (a) A lone ] character is dis-allowed (Perl treats it as data).
77 (b) A back reference to an unmatched subpattern matches an empty string
78 (Perl fails the current match path).
80 (c) A data ] in a character class must be notated as \] because if the
81 first data character in a class is ], it defines an empty class. (In
82 Perl it is not possible to have an empty class.) The empty class []
83 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
84 The negative empty class [^] matches any one character, independently
85 of the DOTALL setting.
87 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
88 non-existent subpattern following a character class starting with ']' and
89 containing () gave an internal compiling error instead of "reference to
90 non-existent subpattern". Fortunately, when the pattern did exist, the
91 compiled code was correct. (When scanning forwards to check for the
92 existencd of the subpattern, it was treating the data ']' as terminating
93 the class, so got the count wrong. When actually compiling, the reference
94 was subsequently set up correctly.)
96 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
97 it was being rejected as not supported by pcre_dfa_exec(), even though
98 other assertions are supported. I have made pcre_dfa_exec() support
99 (*FAIL).
101 16. The implementation of 13c above involved the invention of a new opcode,
102 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
103 cannot be changed at match time, I realized I could make a small
104 improvement to matching performance by compiling OP_ALLANY instead of
105 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
106 on the OP_ANY path.
108 17. Compiling pcretest on Windows with readline support failed without the
109 following two fixes: (1) Make the unistd.h include conditional on
110 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
112 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
113 ncurses library to be included for pcretest when ReadLine support is
114 requested, but also to allow for it to be overridden. This patch came from
115 Daniel Bergström.
117 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
118 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
119 any errors with the current Unicode tables. Thanks to Peter Kankowski for
120 spotting this.
123 Version 7.6 28-Jan-08
124 ---------------------
126 1. A character class containing a very large number of characters with
127 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
128 overflow.
130 2. Patch to cut out the "long long" test in pcrecpp_unittest when
131 HAVE_LONG_LONG is not defined.
133 3. Applied Christian Ehrlicher's patch to update the CMake build files to
134 bring them up to date and include new features. This patch includes:
136 - Fixed PH's badly added libz and libbz2 support.
137 - Fixed a problem with static linking.
138 - Added pcredemo. [But later removed - see 7 below.]
139 - Fixed dftables problem and added an option.
140 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
142 - Added readline support for pcretest.
143 - Added an listing of the option settings after cmake has run.
145 4. A user submitted a patch to Makefile that makes it easy to create
146 "pcre.dll" under mingw when using Configure/Make. I added stuff to
147 Makefile.am that cause it to include this special target, without
148 affecting anything else. Note that the same mingw target plus all
149 the other distribution libraries and programs are now supported
150 when configuring with CMake (see 6 below) instead of with
151 Configure/Make.
153 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
154 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
155 exported in the Windows port". It has not yet been confirmed that the patch
156 solves the problem, but it does no harm.
158 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
159 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
160 with CMake, and also correct the comment about stack recursion.
162 7. Remove the automatic building of pcredemo from the ./configure system and
163 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
164 of a program that users should build themselves after PCRE is installed, so
165 building it automatically is not really right. What is more, it gave
166 trouble in some build environments.
168 8. Further tidies to CMakeLists.txt from Sheri and Christian.
171 Version 7.5 10-Jan-08
172 ---------------------
174 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
175 values in parens when parsing an RE using the C++ wrapper."
177 2. Negative specials like \S did not work in character classes in UTF-8 mode.
178 Characters greater than 255 were excluded from the class instead of being
179 included.
181 3. The same bug as (2) above applied to negated POSIX classes such as
182 [:^space:].
184 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
185 defined or documented. It seems to have been a typo for PCRE_STATIC, so
186 I have changed it.
188 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
189 first named subpattern) and a construct such as (?&a) would reference the
190 first named subpattern whose name started with "a" (in other words, the
191 length check was missing). Both these problems are fixed. "Subpattern name
192 expected" is now given for (?&) (a zero-length name), and this patch also
193 makes it give the same error for \k'' (previously it complained that that
194 was a reference to a non-existent subpattern).
196 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
197 this is right because (?- can be followed by option settings as well as by
198 digits. I have, however, made the messages clearer.
200 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
201 than the number used in the conditional) now cause a compile-time error.
202 This is actually not compatible with Perl, which accepts such patterns, but
203 treats the conditional as always being FALSE (as PCRE used to), but it
204 seems to me that giving a diagnostic is better.
206 8. Change "alphameric" to the more common word "alphanumeric" in comments
207 and messages.
209 9. Fix two occurrences of "backslash" in comments that should have been
210 "backspace".
212 10. Remove two redundant lines of code that can never be obeyed (their function
213 was moved elsewhere).
215 11. The program that makes PCRE's Unicode character property table had a bug
216 which caused it to generate incorrect table entries for sequences of
217 characters that have the same character type, but are in different scripts.
218 It amalgamated them into a single range, with the script of the first of
219 them. In other words, some characters were in the wrong script. There were
220 thirteen such cases, affecting characters in the following ranges:
222 U+002b0 - U+002c1
223 U+0060c - U+0060d
224 U+0061e - U+00612
225 U+0064b - U+0065e
226 U+0074d - U+0076d
227 U+01800 - U+01805
228 U+01d00 - U+01d77
229 U+01d9b - U+01dbf
230 U+0200b - U+0200f
231 U+030fc - U+030fe
232 U+03260 - U+0327f
233 U+0fb46 - U+0fbb1
234 U+10450 - U+1049d
236 12. The -o option (show only the matching part of a line) for pcregrep was not
237 compatible with GNU grep in that, if there was more than one match in a
238 line, it showed only the first of them. It now behaves in the same way as
239 GNU grep.
241 13. If the -o and -v options were combined for pcregrep, it printed a blank
242 line for every non-matching line. GNU grep prints nothing, and pcregrep now
243 does the same. The return code can be used to tell if there were any
244 non-matching lines.
246 14. Added --file-offsets and --line-offsets to pcregrep.
248 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
249 infinitely looping recursion. The bug was that positive lookaheads were not
250 being skipped when checking for a possible empty match (negative lookaheads
251 and both kinds of lookbehind were skipped).
253 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
254 inclusion of <windows.h> to before rather than after the definition of
255 INVALID_FILE_ATTRIBUTES (patch from David Byron).
257 17. Specifying a possessive quantifier with a specific limit for a Unicode
258 character property caused pcre_compile() to compile bad code, which led at
259 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
260 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
261 caused the error; without that there was no problem.
263 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
265 19. Added --enable-pcretest-libreadline.
267 20. In pcrecpp.cc, the variable 'count' was incremented twice in
268 RE::GlobalReplace(). As a result, the number of replacements returned was
269 double what it should be. I removed one of the increments, but Craig sent a
270 later patch that removed the other one (the right fix) and added unit tests
271 that check the return values (which was not done before).
273 21. Several CMake things:
275 (1) Arranged that, when cmake is used on Unix, the libraries end up with
276 the names libpcre and libpcreposix, not just pcre and pcreposix.
278 (2) The above change means that pcretest and pcregrep are now correctly
279 linked with the newly-built libraries, not previously installed ones.
283 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
284 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
285 UTF-8 newline character). The key issue is that the pattern starts .*;
286 this means that the match must be either at the beginning, or after a
287 newline. The bug was in the code for advancing after a failed match and
288 checking that the new position followed a newline. It was not taking
289 account of UTF-8 characters correctly.
291 23. PCRE was behaving differently from Perl in the way it recognized POSIX
292 character classes. PCRE was not treating the sequence [:...:] as a
293 character class unless the ... were all letters. Perl, however, seems to
294 allow any characters between [: and :], though of course it rejects as
295 unknown any "names" that contain non-letters, because all the known class
296 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
297 for example, whereas PCRE did not - it did not recognize a POSIX character
298 class. This seemed a bit dangerous, so the code has been changed to be
299 closer to Perl. The behaviour is not identical to Perl, because PCRE will
300 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
301 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
302 Perl does, and where it didn't before.
304 24. Rewrite so as to remove the single use of %n from pcregrep because in some
305 Windows environments %n is disabled by default.
308 Version 7.4 21-Sep-07
309 ---------------------
311 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
312 means that a class such as [\s] counted as "explicit reference to CR or
313 LF". That isn't really right - the whole point of the change was to try to
314 help when there was an actual mention of one of the two characters. So now
315 the change happens only if \r or \n (or a literal CR or LF) character is
316 encountered.
318 2. The 32-bit options word was also used for 6 internal flags, but the numbers
319 of both had grown to the point where there were only 3 bits left.
320 Fortunately, there was spare space in the data structure, and so I have
321 moved the internal flags into a new 16-bit field to free up more option
322 bits.
324 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
325 but did not set the internal JCHANGED flag - either of these is enough to
326 control the way the "get" function works - but the PCRE_INFO_JCHANGED
327 facility is supposed to tell if (?J) was ever used, so now (?J) at the
328 start sets both bits.
330 4. Added options (at build time, compile time, exec time) to change \R from
331 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
333 5. doc/pcresyntax.html was missing from the distribution.
335 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
336 compatibility, even though it is no longer used.
338 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
339 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
340 windows.h file is present (where different names are used). [This was
341 reversed later after testing - see 16 below.]
343 8. Changed all #include <config.h> to #include "config.h". There were also
344 some further <pcre.h> cases that I changed to "pcre.h".
346 9. When pcregrep was used with the --colour option, it missed the line ending
347 sequence off the lines that it output.
349 10. It was pointed out to me that arrays of string pointers cause lots of
350 relocations when a shared library is dynamically loaded. A technique of
351 using a single long string with a table of offsets can drastically reduce
352 these. I have refactored PCRE in four places to do this. The result is
353 dramatic:
355 Originally: 290
356 After changing UCP table: 187
357 After changing error message table: 43
358 After changing table of "verbs" 36
359 After changing table of Posix names 22
361 Thanks to the folks working on Gregex for glib for this insight.
363 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
364 unicode-properties was also set.
366 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
368 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
369 checked only for CRLF.
371 14. Added casts to pcretest.c to avoid compiler warnings.
373 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
375 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
376 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
377 entirely. This removes changes made in 7 above.
379 17. The CMake files have been updated, and there is now more information about
380 building with CMake in the NON-UNIX-USE document.
383 Version 7.3 28-Aug-07
384 ---------------------
386 1. In the rejigging of the build system that eventually resulted in 7.1, the
387 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
388 brackets there is not right, since it causes compilers to look for an
389 installed pcre.h, not the version that is in the source that is being
390 compiled (which of course may be different). I have changed it back to:
392 #include "pcre.h"
394 I have a vague recollection that the change was concerned with compiling in
395 different directories, but in the new build system, that is taken care of
396 by the VPATH setting the Makefile.
398 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
399 when the subject happened to end in the byte 0x85 (e.g. if the last
400 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
401 characters but of course it shouldn't be taken as a newline when it is part
402 of another character. The bug was that, for an unlimited repeat of . in
403 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
404 characters when looking for a newline.
406 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
408 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
409 in debug output.
411 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
412 long printing in the pcrecpp unittest when running under MinGW.
414 6. ESC_K was left out of the EBCDIC table.
416 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
417 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
418 limit also applies to "virtual nesting" when a pattern is recursive, and in
419 this case 1000 isn't so big. I have been able to remove this limit at the
420 expense of backing off one optimization in certain circumstances. Normally,
421 when pcre_exec() would call its internal match() function recursively and
422 immediately return the result unconditionally, it uses a "tail recursion"
423 feature to save stack. However, when a subpattern that can match an empty
424 string has an unlimited repetition quantifier, it no longer makes this
425 optimization. That gives it a stack frame in which to save the data for
426 checking that an empty string has been matched. Previously this was taken
427 from the 1000-entry workspace that had been reserved. So now there is no
428 explicit limit, but more stack is used.
430 8. Applied Daniel's patches to solve problems with the import/export magic
431 syntax that is required for Windows, and which was going wrong for the
432 pcreposix and pcrecpp parts of the library. These were overlooked when this
433 problem was solved for the main library.
435 9. There were some crude static tests to avoid integer overflow when computing
436 the size of patterns that contain repeated groups with explicit upper
437 limits. As the maximum quantifier is 65535, the maximum group length was
438 set at 30,000 so that the product of these two numbers did not overflow a
439 32-bit integer. However, it turns out that people want to use groups that
440 are longer than 30,000 bytes (though not repeat them that many times).
441 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
442 made it possible to implement the integer overflow checks in a much more
443 dynamic way, which I have now done. The artificial limitation on group
444 length has been removed - we now have only the limit on the total length of
445 the compiled pattern, which depends on the LINK_SIZE setting.
447 10. Fixed a bug in the documentation for get/copy named substring when
448 duplicate names are permitted. If none of the named substrings are set, the
449 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
450 empty string.
452 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
453 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
454 because the ] is interpreted as the first data character and the
455 terminating ] is not found. PCRE has been made compatible with Perl in this
456 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
457 cause memory overwriting.
459 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
460 string has been matched (to stop an infinite loop). It was not recognizing
461 a conditional subpattern that could match an empty string if that
462 subpattern was within another subpattern. For example, it looped when
463 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
464 condition was not nested. This bug has been fixed.
466 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
467 past the start of the subject in the presence of bytes with the top bit
468 set, for example "\x8aBCD".
470 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
471 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
473 14. Optimized (?!) to (*FAIL).
475 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
476 This restricts code points to be within the range 0 to 0x10FFFF, excluding
477 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
478 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
479 does: it's just the validity check that is more restrictive.
481 16. Inserted checks for integer overflows during escape sequence (backslash)
482 processing, and also fixed erroneous offset values for syntax errors during
483 backslash processing.
485 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
486 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
488 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
489 caused an overrun.
491 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
492 something other than just ASCII characters) inside a group that had an
493 unlimited repeat caused a loop at compile time (while checking to see
494 whether the group could match an empty string).
496 20. Debugging a pattern containing \p or \P could cause a crash. For example,
497 [\P{Any}] did so. (Error in the code for printing property names.)
499 21. An orphan \E inside a character class could cause a crash.
501 22. A repeated capturing bracket such as (A)? could cause a wild memory
502 reference during compilation.
504 23. There are several functions in pcre_compile() that scan along a compiled
505 expression for various reasons (e.g. to see if it's fixed length for look
506 behind). There were bugs in these functions when a repeated \p or \P was
507 present in the pattern. These operators have additional parameters compared
508 with \d, etc, and these were not being taken into account when moving along
509 the compiled data. Specifically:
511 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
512 length.
514 (b) An item such as \pL+ within a repeated group could cause crashes or
515 loops.
517 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
518 "reference to non-existent subpattern" error.
520 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
522 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
523 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
525 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
527 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
528 character were causing crashes (broken optimization).
530 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
531 \p or \P) caused a compile-time loop.
533 28. More problems have arisen in unanchored patterns when CRLF is a valid line
534 break. For example, the unstudied pattern [\r\n]A does not match the string
535 "\r\nA" because change 7.0/46 below moves the current point on by two
536 characters after failing to match at the start. However, the pattern \nA
537 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
538 the same is true. There doesn't seem any very clean way out of this, but
539 what I have chosen to do makes the common cases work: PCRE now takes note
540 of whether there can be an explicit match for \r or \n anywhere in the
541 pattern, and if so, 7.0/46 no longer applies. As part of this change,
542 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
543 pattern has explicit CR or LF references.
545 29. Added (*CR) etc for changing newline setting at start of pattern.
548 Version 7.2 19-Jun-07
549 ---------------------
551 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
552 which is apparently normally available under Windows.
554 2. Re-jig the pcregrep tests with different newline settings in an attempt
555 to make them independent of the local environment's newline setting.
557 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
559 4. Some of the "internals" tests were previously cut out when the link size
560 was not 2, because the output contained actual offsets. The recent new
561 "Z" feature of pcretest means that these can be cut out, making the tests
562 usable with all link sizes.
564 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
565 stack recursion. This gives a massive performance boost under BSD, but just
566 a small improvement under Linux. However, it saves one field in the frame
567 in all cases.
569 6. Added more features from the forthcoming Perl 5.10:
571 (a) (?-n) (where n is a string of digits) is a relative subroutine or
572 recursion call. It refers to the nth most recently opened parentheses.
574 (b) (?+n) is also a relative subroutine call; it refers to the nth next
575 to be opened parentheses.
577 (c) Conditions that refer to capturing parentheses can be specified
578 relatively, for example, (?(-2)... or (?(+3)...
580 (d) \K resets the start of the current match so that everything before
581 is not part of it.
583 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
585 (f) \g{name} is another synonym - part of Perl 5.10's unification of
586 reference syntax.
588 (g) (?| introduces a group in which the numbering of parentheses in each
589 alternative starts with the same number.
591 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
593 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
596 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
597 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
598 for detecting groups that can match an empty string.
600 9. A pattern with a very large number of alternatives (more than several
601 hundred) was running out of internal workspace during the pre-compile
602 phase, where pcre_compile() figures out how much memory will be needed. A
603 bit of new cunning has reduced the workspace needed for groups with
604 alternatives. The 1000-alternative test pattern now uses 12 bytes of
605 workspace instead of running out of the 4096 that are available.
607 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
609 11. Applied patch from Google to remove an optimization that didn't quite work.
610 The report of the bug said:
612 pcrecpp::RE("a*").FullMatch("aaa") matches, while
613 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
614 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
616 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
617 it matched the wrong number of bytes.
620 Version 7.1 24-Apr-07
621 ---------------------
623 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
624 that is more "standard", making use of automake and other Autotools. There
625 is some re-arrangement of the files and adjustment of comments consequent
626 on this.
628 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
629 for recursive directory scanning broke on some systems because the files
630 are not scanned in any specific order and on different systems the order
631 was different. A call to "sort" has been inserted into RunGrepTest for the
632 approprate test as a short-term fix. In the longer term there may be an
633 alternative.
635 3. I had an email from Eric Raymond about problems translating some of PCRE's
636 man pages to HTML (despite the fact that I distribute HTML pages, some
637 people do their own conversions for various reasons). The problems
638 concerned the use of low-level troff macros .br and .in. I have therefore
639 removed all such uses from the man pages (some were redundant, some could
640 be replaced by .nf/.fi pairs). The 132html script that I use to generate
641 HTML has been updated to handle .nf/.fi and to complain if it encounters
642 .br or .in.
644 4. Updated comments in configure.ac that get placed in config.h.in and also
645 arranged for config.h to be included in the distribution, with the name
646 config.h.generic, for the benefit of those who have to compile without
647 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
649 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
650 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
651 makevp.bat for the current PCRE, using the additional files
652 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
654 6. A Windows user reported a minor discrepancy with test 2, which turned out
655 to be caused by a trailing space on an input line that had got lost in his
656 copy. The trailing space was an accident, so I've just removed it.
658 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
659 that is needed.
661 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
662 as "const" (a) because they are and (b) because it helps the PHP
663 maintainers who have recently made a script to detect big data structures
664 in the php code that should be moved to the .rodata section. I remembered
665 to update Builducptable as well, so it won't revert if ucptable.h is ever
666 re-created.
668 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
669 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
670 order to be able to cut out the UTF-8 tables in the latter when UTF-8
671 support is not required. This saves 1.5-2K of code, which is important in
672 some applications.
674 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
675 so as not to refer to the tables, even though these functions will never be
676 called when UTF-8 support is disabled. Otherwise there are problems with a
677 shared library.
679 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
681 (a) It was defining its arguments as char * instead of void *.
683 (b) It was assuming that all moves were upwards in memory; this was true
684 a long time ago when I wrote it, but is no longer the case.
686 The emulated memove() is provided for those environments that have neither
687 memmove() nor bcopy(). I didn't think anyone used it these days, but that
688 is clearly not the case, as these two bugs were recently reported.
690 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
691 and Detrail to create the HTML documentation, the .txt form of the man
692 pages, and it removes trailing spaces from listed files. It also creates
693 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
694 case, it wraps all the #defines with #ifndefs. This script should be run
695 before "make dist".
697 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
698 with Unicode property support.
700 (a) For a maximizing quantifier, if the two different cases of the
701 character were of different lengths in their UTF-8 codings (there are
702 some cases like this - I found 11), and the matching function had to
703 back up over a mixture of the two cases, it incorrectly assumed they
704 were both the same length.
706 (b) When PCRE was configured to use the heap rather than the stack for
707 recursion during matching, it was not correctly preserving the data for
708 the other case of a UTF-8 character when checking ahead for a match
709 while processing a minimizing repeat. If the check also involved
710 matching a wide character, but failed, corruption could cause an
711 erroneous result when trying to check for a repeat of the original
712 character.
714 13. Some tidying changes to the testing mechanism:
716 (a) The RunTest script now detects the internal link size and whether there
717 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
718 values substituted by "configure". (The RunGrepTest script already did
719 this for UTF-8.) The configure.ac script no longer substitutes the
720 relevant variables.
722 (b) The debugging options /B and /D in pcretest show the compiled bytecode
723 with length and offset values. This means that the output is different
724 for different internal link sizes. Test 2 is skipped for link sizes
725 other than 2 because of this, bypassing the problem. Unfortunately,
726 there was also a test in test 3 (the locale tests) that used /B and
727 failed for link sizes other than 2. Rather than cut the whole test out,
728 I have added a new /Z option to pcretest that replaces the length and
729 offset values with spaces. This is now used to make test 3 independent
730 of link size. (Test 2 will be tidied up later.)
732 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
733 segmentation fault instead of returning the appropriate error message.
735 15. In multiline mode when the newline sequence was set to "any", the pattern
736 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
737 This doesn't seem right; it now treats the CRLF combination as the line
738 ending, and so does not match in that case. It's only a pattern such as ^$
739 that would hit this one: something like ^ABC$ would have failed after \r
740 and then tried again after \r\n.
742 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
743 in an attempt to make files that differ only in their line terminators
744 compare equal. This works on Linux.
746 17. Under certain error circumstances pcregrep might try to free random memory
747 as it exited. This is now fixed, thanks to valgrind.
749 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
750 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
751 was because its rules for how to advance for /g after matching an empty
752 string at the end of a line did not allow for this case. They now check for
753 it specially.
755 20. pcretest is supposed to handle patterns and data of any length, by
756 extending its buffers when necessary. It was getting this wrong when the
757 buffer for a data line had to be extended.
759 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
760 CRLF as a newline sequence.
762 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
763 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
764 I have nevertheless tidied it up.
766 23. Added some casts to kill warnings from HP-UX ia64 compiler.
768 24. Added a man page for pcre-config.
771 Version 7.0 19-Dec-06
772 ---------------------
774 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
775 moving to gcc 4.1.1.
777 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
778 sys/time.h, which is documented as needed for this function. It doesn't
779 seem to matter on Linux, but it showed up on some releases of OS X.
781 3. It seems that there are systems where bytes whose values are greater than
782 127 match isprint() in the "C" locale. The "C" locale should be the
783 default when a C program starts up. In most systems, only ASCII printing
784 characters match isprint(). This difference caused the output from pcretest
785 to vary, making some of the tests fail. I have changed pcretest so that:
787 (a) When it is outputting text in the compiled version of a pattern, bytes
788 other than 32-126 are always shown as hex escapes.
790 (b) When it is outputting text that is a matched part of a subject string,
791 it does the same, unless a different locale has been set for the match
792 (using the /L modifier). In this case, it uses isprint() to decide.
794 4. Fixed a major bug that caused incorrect computation of the amount of memory
795 required for a compiled pattern when options that changed within the
796 pattern affected the logic of the preliminary scan that determines the
797 length. The relevant options are -x, and -i in UTF-8 mode. The result was
798 that the computed length was too small. The symptoms of this bug were
799 either the PCRE error "internal error: code overflow" from pcre_compile(),
800 or a glibc crash with a message such as "pcretest: free(): invalid next
801 size (fast)". Examples of patterns that provoked this bug (shown in
802 pcretest format) are:
804 /(?-x: )/x
805 /(?x)(?-x: \s*#\s*)/
806 /((?i)[\x{c0}])/8
807 /(?i:[\x{c0}])/8
809 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
810 is now done differently.
812 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
813 wrapper classes; (b) implement a new function in the C++ scanner that is
814 more efficient than the old way of doing things because it avoids levels of
815 recursion in the regex matching; (c) add a paragraph to the documentation
816 for the FullMatch() function.
818 6. The escape sequence \n was being treated as whatever was defined as
819 "newline". Not only was this contrary to the documentation, which states
820 that \n is character 10 (hex 0A), but it also went horribly wrong when
821 "newline" was defined as CRLF. This has been fixed.
823 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
824 was being set to -1 for the "end of line" case (supposedly a value that no
825 character can have). Though this value is never used (the check for end of
826 line is "zero bytes in current character"), it caused compiler complaints.
827 I've changed it to 0xffffffff.
829 8. In pcre_version.c, the version string was being built by a sequence of
830 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
831 string (as it is for production releases) called a macro with an empty
832 argument. The C standard says the result of this is undefined. The gcc
833 compiler treats it as an empty string (which was what was wanted) but it is
834 reported that Visual C gives an error. The source has been hacked around to
835 avoid this problem.
837 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
838 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
839 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
840 of them did).
842 10. Originally, pcretest opened its input and output without "b"; then I was
843 told that "b" was needed in some environments, so it was added for release
844 5.0 to both the input and output. (It makes no difference on Unix-like
845 systems.) Later I was told that it is wrong for the input on Windows. I've
846 now abstracted the modes into two macros, to make it easier to fiddle with
847 them, and removed "b" from the input mode under Windows.
849 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
851 12. Added -help and --help to pcretest as an official way of being reminded
852 of the options.
854 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
855 and pcrecpp.cc because they annoy compilers at high warning levels.
857 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
859 15. Fixed an occurrence of == in configure.ac that should have been = (shell
860 scripts are not C programs :-) and which was not noticed because it works
861 on Linux.
863 16. pcretest is supposed to handle any length of pattern and data line (as one
864 line or as a continued sequence of lines) by extending its input buffer if
865 necessary. This feature was broken for very long pattern lines, leading to
866 a string of junk being passed to pcre_compile() if the pattern was longer
867 than about 50K.
869 17. I have done a major re-factoring of the way pcre_compile() computes the
870 amount of memory needed for a compiled pattern. Previously, there was code
871 that made a preliminary scan of the pattern in order to do this. That was
872 OK when PCRE was new, but as the facilities have expanded, it has become
873 harder and harder to keep it in step with the real compile phase, and there
874 have been a number of bugs (see for example, 4 above). I have now found a
875 cunning way of running the real compile function in a "fake" mode that
876 enables it to compute how much memory it would need, while actually only
877 ever using a few hundred bytes of working memory and without too many
878 tests of the mode. This should make future maintenance and development
879 easier. A side effect of this work is that the limit of 200 on the nesting
880 depth of parentheses has been removed (though this was never a serious
881 limitation, I suspect). However, there is a downside: pcre_compile() now
882 runs more slowly than before (30% or more, depending on the pattern). I
883 hope this isn't a big issue. There is no effect on runtime performance.
885 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
886 newline (only possible for the last line of a file) and it was a
887 pattern that set a locale (followed by /Lsomething), pcretest crashed.
889 19. Added additional timing features to pcretest. (1) The -tm option now times
890 matching only, not compiling. (2) Both -t and -tm can be followed, as a
891 separate command line item, by a number that specifies the number of
892 repeats to use when timing. The default is 50000; this gives better
893 precision, but takes uncomfortably long for very large patterns.
895 20. Extended pcre_study() to be more clever in cases where a branch of a
896 subpattern has no definite first character. For example, (a*|b*)[cd] would
897 previously give no result from pcre_study(). Now it recognizes that the
898 first character must be a, b, c, or d.
900 21. There was an incorrect error "recursive call could loop indefinitely" if
901 a subpattern (or the entire pattern) that was being tested for matching an
902 empty string contained only one non-empty item after a nested subpattern.
903 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
904 incorrectly, because the \d was being skipped in the check.
906 22. The pcretest program now has a new pattern option /B and a command line
907 option -b, which is equivalent to adding /B to every pattern. This causes
908 it to show the compiled bytecode, without the additional information that
909 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
910 is the same as /B/I).
912 23. A new optimization is now able automatically to treat some sequences such
913 as a*b as a*+b. More specifically, if something simple (such as a character
914 or a simple class like \d) has an unlimited quantifier, and is followed by
915 something that cannot possibly match the quantified thing, the quantifier
916 is automatically "possessified".
918 24. A recursive reference to a subpattern whose number was greater than 39
919 went wrong under certain circumstances in UTF-8 mode. This bug could also
920 have affected the operation of pcre_study().
922 25. Realized that a little bit of performance could be had by replacing
923 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
925 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
927 27. Possessive quantifiers such as a++ were previously implemented by turning
928 them into atomic groups such as ($>a+). Now they have their own opcodes,
929 which improves performance. This includes the automatically created ones
930 from 23 above.
932 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
933 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
934 the first matched character to be a colon. This applied both to named and
935 numbered groups.
937 29. The ucpinternal.h header file was missing its idempotency #ifdef.
939 30. I was sent a "project" file called libpcre.a.dev which I understand makes
940 building PCRE on Windows easier, so I have included it in the distribution.
942 31. There is now a check in pcretest against a ridiculously large number being
943 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
944 loop, the loop is abandoned.
946 32. Forward references to subpatterns in conditions such as (?(2)...) where
947 subpattern 2 is defined later cause pcre_compile() to search forwards in
948 the pattern for the relevant set of parentheses. This search went wrong
949 when there were unescaped parentheses in a character class, parentheses
950 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
952 33. "Subroutine" calls and backreferences were previously restricted to
953 referencing subpatterns earlier in the regex. This restriction has now
954 been removed.
956 34. Added a number of extra features that are going to be in Perl 5.10. On the
957 whole, these are just syntactic alternatives for features that PCRE had
958 previously implemented using the Python syntax or my own invention. The
959 other formats are all retained for compatibility.
961 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
962 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
963 also .NET compatible.
965 (b) A recursion or subroutine call to a named group can now be defined as
966 (?&name) as well as (?P>name).
968 (c) A backreference to a named group can now be defined as \k<name> or
969 \k'name' as well as (?P=name). The new forms, as well as being in Perl
970 5.10, are also .NET compatible.
972 (d) A conditional reference to a named group can now use the syntax
973 (?(<name>) or (?('name') as well as (?(name).
975 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
976 groups (named and numbered) that are never evaluated inline, but can be
977 called as "subroutines" from elsewhere. In effect, the DEFINE condition
978 is always false. There may be only one alternative in such a group.
980 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
981 as the simple (?(R). The condition is true only if the most recent
982 recursion is that of the given number or name. It does not search out
983 through the entire recursion stack.
985 (g) The escape \gN or \g{N} has been added, where N is a positive or
986 negative number, specifying an absolute or relative reference.
988 35. Tidied to get rid of some further signed/unsigned compiler warnings and
989 some "unreachable code" warnings.
991 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
992 things, this adds five new scripts.
994 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
995 There were also incompatibilities regarding the handling of \Q..\E inside
996 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
997 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
999 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1000 matches an empty string, and forcibly breaks the loop. There were bugs in
1001 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1002 against aaaa the result was just "a" rather than "aaaa", for example. Two
1003 separate and independent bugs (that affected different cases) have been
1004 fixed.
1006 39. Refactored the code to abolish the use of different opcodes for small
1007 capturing bracket numbers. This is a tidy that I avoided doing when I
1008 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1009 The new approach is not only tidier, it makes it possible to reduce the
1010 memory needed to fix the previous bug (38).
1012 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1013 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1014 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1015 mode.
1017 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1018 report.
1020 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1021 copy construction and assignment in the C++ wrapper.
1023 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1024 couple of bugs that could have given wrong results in the "--newline=crlf"
1025 case.
1027 44. Added a number of casts and did some reorganization of signed/unsigned int
1028 variables following suggestions from Dair Grant. Also renamed the variable
1029 "this" as "item" because it is a C++ keyword.
1031 45. Arranged for dftables to add
1033 #include "pcre_internal.h"
1035 to pcre_chartables.c because without it, gcc 4.x may remove the array
1036 definition from the final binary if PCRE is built into a static library and
1037 dead code stripping is activated.
1039 46. For an unanchored pattern, if a match attempt fails at the start of a
1040 newline sequence, and the newline setting is CRLF or ANY, and the next two
1041 characters are CRLF, advance by two characters instead of one.
1044 Version 6.7 04-Jul-06
1045 ---------------------
1047 1. In order to handle tests when input lines are enormously long, pcretest has
1048 been re-factored so that it automatically extends its buffers when
1049 necessary. The code is crude, but this _is_ just a test program. The
1050 default size has been increased from 32K to 50K.
1052 2. The code in pcre_study() was using the value of the re argument before
1053 testing it for NULL. (Of course, in any sensible call of the function, it
1054 won't be NULL.)
1056 3. The memmove() emulation function in pcre_internal.h, which is used on
1057 systems that lack both memmove() and bcopy() - that is, hardly ever -
1058 was missing a "static" storage class specifier.
1060 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1061 containing an extended class (one that cannot be represented by a bitmap
1062 because it contains high-valued characters or Unicode property items, e.g.
1063 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1064 pattern, but PCRE should not loop if you do not (it no longer does).
1065 [Detail: two cases were found: (a) a repeated subpattern containing an
1066 extended class; (b) a recursive reference to a subpattern that followed a
1067 previous extended class. It wasn't skipping over the extended class
1068 correctly when UTF-8 mode was not set.]
1070 5. A negated single-character class was not being recognized as fixed-length
1071 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1072 compile error "lookbehind assertion is not fixed length".
1074 6. The RunPerlTest auxiliary script was showing an unexpected difference
1075 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1076 write a Perl script that can interpret lines of an input file either as
1077 byte characters or as UTF-8, which is what "perltest" was being required to
1078 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1079 can't do is switch easily at run time between having the "use utf8;" pragma
1080 or not. In the end, I fudged it by using the RunPerlTest script to insert
1081 "use utf8;" explicitly for the UTF-8 tests.
1083 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1084 the end of the subject string, contrary to the documentation and to what
1085 Perl does. This was true of both matching functions. Now it matches only at
1086 the start of the subject and immediately after *internal* newlines.
1088 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1089 a pointer to an int instead of a pointer to an unsigned long int. This
1090 caused problems on 64-bit systems.
1092 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1093 instance of the 'standard' template library not being so standard".
1095 10. There was no check on the number of named subpatterns nor the maximum
1096 length of a subpattern name. The product of these values is used to compute
1097 the size of the memory block for a compiled pattern. By supplying a very
1098 long subpattern name and a large number of named subpatterns, the size
1099 computation could be caused to overflow. This is now prevented by limiting
1100 the length of names to 32 characters, and the number of named subpatterns
1101 to 10,000.
1103 11. Subpatterns that are repeated with specific counts have to be replicated in
1104 the compiled pattern. The size of memory for this was computed from the
1105 length of the subpattern and the repeat count. The latter is limited to
1106 65535, but there was no limit on the former, meaning that integer overflow
1107 could in principle occur. The compiled length of a repeated subpattern is
1108 now limited to 30,000 bytes in order to prevent this.
1110 12. Added the optional facility to have named substrings with the same name.
1112 13. Added the ability to use a named substring as a condition, using the
1113 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1114 are numbers (not recommended). Forward references are permitted.
1116 14. Added forward references in named backreferences (if you see what I mean).
1118 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1119 pattern could run off the end of the subject. For example, the pattern
1120 "(?s)(.{1,5})"8 did this with the subject "ab".
1122 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1123 PCRE_CASELESS was set when matching characters that were quantified with ?
1124 or *.
1126 17. A character class other than a single negated character that had a minimum
1127 but no maximum quantifier - for example [ab]{6,} - was not handled
1128 correctly by pce_dfa_exec(). It would match only one character.
1130 18. A valid (though odd) pattern that looked like a POSIX character
1131 class but used an invalid character after [ (for example [[,abc,]]) caused
1132 pcre_compile() to give the error "Failed: internal error: code overflow" or
1133 in some cases to crash with a glibc free() error. This could even happen if
1134 the pattern terminated after [[ but there just happened to be a sequence of
1135 letters, a binary zero, and a closing ] in the memory that followed.
1137 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1138 over the years. Originally (before any Unicode support), just the bottom 8
1139 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1140 output from "man perlunicode" includes this:
1142 The regular expression compiler produces polymorphic opcodes. That
1143 is, the pattern adapts to the data and automatically switches to
1144 the Unicode character scheme when presented with Unicode data--or
1145 instead uses a traditional byte scheme when presented with byte
1146 data.
1148 Sadly, a wide octal escape does not cause a switch, and in a string with
1149 no other multibyte characters, these octal escapes are treated as before.
1150 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1151 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1152 Unicode string.
1154 I have not perpetrated such confusion in PCRE. Up till now, it took just
1155 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1156 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1157 translate to the appropriate multibyte character.
1159 29. Applied some refactoring to reduce the number of warnings from Microsoft
1160 and Borland compilers. This has included removing the fudge introduced
1161 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1162 a warning about an unused variable.
1164 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1165 characters since release 4.0, because Perl (from release 5.004) does not.
1166 [Or at least, is documented not to: some releases seem to be in conflict
1167 with the documentation.] However, when a pattern was studied with
1168 pcre_study() and all its branches started with \s, PCRE still included VT
1169 as a possible starting character. Of course, this did no harm; it just
1170 caused an unnecessary match attempt.
1172 22. Removed a now-redundant internal flag bit that recorded the fact that case
1173 dependency changed within the pattern. This was once needed for "required
1174 byte" processing, but is no longer used. This recovers a now-scarce options
1175 bit. Also moved the least significant internal flag bit to the most-
1176 significant bit of the word, which was not previously used (hangover from
1177 the days when it was an int rather than a uint) to free up another bit for
1178 the future.
1180 23. Added support for CRLF line endings as well as CR and LF. As well as the
1181 default being selectable at build time, it can now be changed at runtime
1182 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1183 specify that it is scanning data with non-default line endings.
1185 24. Changed the definition of CXXLINK to make it agree with the definition of
1186 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1188 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1189 recursions. This makes a big different to stack usage for some patterns.
1191 26. If a subpattern containing a named recursion or subroutine reference such
1192 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1193 the space required for the compiled pattern went wrong and gave too small a
1194 value. Depending on the environment, this could lead to "Failed: internal
1195 error: code overflow at offset 49" or "glibc detected double free or
1196 corruption" errors.
1198 27. Applied patches from Google (a) to support the new newline modes and (b) to
1199 advance over multibyte UTF-8 characters in GlobalReplace.
1201 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1202 difference for some implementation of PCRE in some Windows version.
1204 29. Added some extra testing facilities to pcretest:
1206 \q<number> in a data line sets the "match limit" value
1207 \Q<number> in a data line sets the "match recursion limt" value
1208 -S <number> sets the stack size, where <number> is in megabytes
1210 The -S option isn't available for Windows.
1213 Version 6.6 06-Feb-06
1214 ---------------------
1216 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1217 in pcreposix.h. I have copied the definition from pcre.h.
1219 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1220 because pcre.h is no longer a built file.
1222 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1223 not normally included in the compiled code.
1226 Version 6.5 01-Feb-06
1227 ---------------------
1229 1. When using the partial match feature with pcre_dfa_exec(), it was not
1230 anchoring the second and subsequent partial matches at the new starting
1231 point. This could lead to incorrect results. For example, with the pattern
1232 /1234/, partially matching against "123" and then "a4" gave a match.
1234 2. Changes to pcregrep:
1236 (a) All non-match returns from pcre_exec() were being treated as failures
1237 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1238 error message is output. Some extra information is given for the
1240 probably the only errors that are likely to be caused by users (by
1241 specifying a regex that has nested indefinite repeats, for instance).
1242 If there are more than 20 of these errors, pcregrep is abandoned.
1244 (b) A binary zero was treated as data while matching, but terminated the
1245 output line if it was written out. This has been fixed: binary zeroes
1246 are now no different to any other data bytes.
1248 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1249 used to set a locale for matching. The --locale=xxxx long option has
1250 been added (no short equivalent) to specify a locale explicitly on the
1251 pcregrep command, overriding the environment variables.
1253 (d) When -B was used with -n, some line numbers in the output were one less
1254 than they should have been.
1256 (e) Added the -o (--only-matching) option.
1258 (f) If -A or -C was used with -c (count only), some lines of context were
1259 accidentally printed for the final match.
1261 (g) Added the -H (--with-filename) option.
1263 (h) The combination of options -rh failed to suppress file names for files
1264 that were found from directory arguments.
1266 (i) Added the -D (--devices) and -d (--directories) options.
1268 (j) Added the -F (--fixed-strings) option.
1270 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1272 (l) Added the --colo(u)r option.
1274 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1275 is not present by default.
1277 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1278 items such as (?R) or (?1), when the recursion could match a number of
1279 alternatives. If it matched one of the alternatives, but subsequently,
1280 outside the recursion, there was a failure, the code tried to back up into
1281 the recursion. However, because of the way PCRE is implemented, this is not
1282 possible, and the result was an incorrect result from the match.
1284 In order to prevent this happening, the specification of recursion has
1285 been changed so that all such subpatterns are automatically treated as
1286 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1288 4. I had overlooked the fact that, in some locales, there are characters for
1289 which isalpha() is true but neither isupper() nor islower() are true. In
1290 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1291 and ordfeminine) are like this. This affected the treatment of \w and \W
1292 when they appeared in character classes, but not when they appeared outside
1293 a character class. The bit map for "word" characters is now created
1294 separately from the results of isalnum() instead of just taking it from the
1295 upper, lower, and digit maps. (Plus the underscore character, of course.)
1297 5. The above bug also affected the handling of POSIX character classes such as
1298 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1299 permanent tables. Instead, the bit maps for such a class were previously
1300 created as the appropriate unions of the upper, lower, and digit bitmaps.
1301 Now they are created by subtraction from the [[:word:]] class, which has
1302 its own bitmap.
1304 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1305 It is created by subtracting the vertical space characters (\x09, \x0a,
1306 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1307 subtraction was done in the overall bitmap for a character class, meaning
1308 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1309 be recognized. This bug has been fixed.
1311 7. Patches from the folks at Google:
1313 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1314 real life, but is still worth protecting against".
1316 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1317 regular expressions".
1319 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1320 have it.
1322 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1323 "configure" and the latter not, in order to fix a problem somebody had
1324 with compiling the Arg class on HP-UX.
1326 (e) Improve the error-handling of the C++ wrapper a little bit.
1328 (f) New tests for checking recursion limiting.
1330 8. The pcre_memmove() function, which is used only if the environment does not
1331 have a standard memmove() function (and is therefore rarely compiled),
1332 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1333 returning a result (though PCRE never actually uses the result).
1335 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1336 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1337 returned instead of calling malloc() with an overflowing number that would
1338 most likely cause subsequent chaos.
1340 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1342 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1343 with this option is matched, the nmatch and pmatch options of regexec() are
1344 ignored.
1346 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1347 provided in case anyone wants to the the POSIX interface with UTF-8
1348 strings.
1350 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1351 C++ linking (needed for some HP-UX environments).
1353 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1354 (unused parameter) and in the pcre_printint() function (omitted "default"
1355 switch label when the default is to do nothing).
1357 15. Added some code to make it possible, when PCRE is compiled as a C++
1358 library, to replace subject pointers for pcre_exec() with a smart pointer
1359 class, thus making it possible to process discontinuous strings.
1361 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1362 much the same function. They were added by different people who were trying
1363 to make PCRE easy to compile on non-Unix systems. It has been suggested
1364 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1365 for compiling on Windows systems. I have therefore replaced it with
1366 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1367 defaults to "extern" for C or "extern C" for C++, which works fine on
1368 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1369 SCOPE with something explicit in config.h. In addition:
1371 (a) pcreposix.h still had just "extern" instead of either of these macros;
1372 I have replaced it with PCRE_DATA_SCOPE.
1374 (b) Functions such as _pcre_xclass(), which are internal to the library,
1375 but external in the C sense, all had PCRE_EXPORT in their definitions.
1376 This is apparently wrong for the Windows case, so I have removed it.
1377 (It makes no difference on Unix-like systems.)
1379 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1380 of recursive calls to match(). This is different to MATCH_LIMIT because
1381 that limits the total number of calls to match(), not all of which increase
1382 the depth of recursion. Limiting the recursion depth limits the amount of
1383 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1384 when PCRE is compiled, and changed at run time. A patch from Google adds
1385 this functionality to the C++ interface.
1387 18. Changes to the handling of Unicode character properties:
1389 (a) Updated the table to Unicode 4.1.0.
1391 (b) Recognize characters that are not in the table as "Cn" (undefined).
1393 (c) I revised the way the table is implemented to a much improved format
1394 which includes recognition of ranges. It now supports the ranges that
1395 are defined in UnicodeData.txt, and it also amalgamates other
1396 characters into ranges. This has reduced the number of entries in the
1397 table from around 16,000 to around 3,000, thus reducing its size
1398 considerably. I realized I did not need to use a tree structure after
1399 all - a binary chop search is just as efficient. Having reduced the
1400 number of entries, I extended their size from 6 bytes to 8 bytes to
1401 allow for more data.
1403 (d) Added support for Unicode script names via properties such as \p{Han}.
1405 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1406 matching that character.
1408 20. When matching a repeated Unicode property with a minimum greater than zero,
1409 (for example \pL{2,}), PCRE could look past the end of the subject if it
1410 reached it while seeking the minimum number of characters. This could
1411 happen only if some of the characters were more than one byte long, because
1412 there is a check for at least the minimum number of bytes.
1414 21. Refactored the implementation of \p and \P so as to be more general, to
1415 allow for more different types of property in future. This has changed the
1416 compiled form incompatibly. Anybody with saved compiled patterns that use
1417 \p or \P will have to recompile them.
1419 22. Added "Any" and "L&" to the supported property types.
1421 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1422 but give a compile time error if the value is greater than 0xff.
1424 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1425 accidentally not being installed or uninstalled.
1427 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1428 made were to insert the current release number. This seemed silly, because
1429 it made things harder for people building PCRE on systems that don't run
1430 "configure". I have turned pcre.h into a distributed file, no longer built
1431 by "configure", with the version identification directly included. There is
1432 no longer a pcre.h.in file.
1434 However, this change necessitated a change to the pcre-config script as
1435 well. It is built from pcre-config.in, and one of the substitutions was the
1436 release number. I have updated configure.ac so that ./configure now finds
1437 the release number by grepping pcre.h.
1439 26. Added the ability to run the tests under valgrind.
1442 Version 6.4 05-Sep-05
1443 ---------------------
1445 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1446 "--" to be printed when multiple files were scanned, even when none of the
1447 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1448 consider it to be a bug, and have restored the previous behaviour.
1450 2. A couple of code tidies to get rid of compiler warnings.
1452 3. The pcretest program used to cheat by referring to symbols in the library
1453 whose names begin with _pcre_. These are internal symbols that are not
1454 really supposed to be visible externally, and in some environments it is
1455 possible to suppress them. The cheating is now confined to including
1456 certain files from the library's source, which is a bit cleaner.
1458 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1459 file's purpose clearer.
1461 5. Reorganized pcre_ucp_findchar().
1464 Version 6.3 15-Aug-05
1465 ---------------------
1467 1. The file libpcre.pc.in did not have general read permission in the tarball.
1469 2. There were some problems when building without C++ support:
1471 (a) If C++ support was not built, "make install" and "make test" still
1472 tried to test it.
1474 (b) There were problems when the value of CXX was explicitly set. Some
1475 changes have been made to try to fix these, and ...
1477 (c) --disable-cpp can now be used to explicitly disable C++ support.
1479 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1480 backslash in a target when C++ was disabled. This confuses some
1481 versions of "make", apparently. Using an intermediate variable solves
1482 this. (Same for CPP_LOBJ.)
1484 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1485 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1486 necessary on certain architectures.
1488 4. Added a setting of -export-symbols-regex to the link command to remove
1489 those symbols that are exported in the C sense, but actually are local
1490 within the library, and not documented. Their names all begin with
1491 "_pcre_". This is not a perfect job, because (a) we have to except some
1492 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1493 available (and never for static libraries). I have made a note to try to
1494 find a way round (a) in the future.
1497 Version 6.2 01-Aug-05
1498 ---------------------
1500 1. There was no test for integer overflow of quantifier values. A construction
1501 such as {1111111111111111} would give undefined results. What is worse, if
1502 a minimum quantifier for a parenthesized subpattern overflowed and became
1503 negative, the calculation of the memory size went wrong. This could have
1504 led to memory overwriting.
1506 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1508 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1509 operating environments where this matters.
1511 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1512 PCRE options from within the C++ wrapper.
1514 5. Named capturing subpatterns were not being correctly counted when a pattern
1515 was compiled. This caused two problems: (a) If there were more than 100
1516 such subpatterns, the calculation of the memory needed for the whole
1517 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1518 back references of the form \12, where the number was greater than 9, were
1519 not recognized as back references, even though there were sufficient
1520 previous subpatterns.
1522 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1523 versions of gcc, e.g. 2.95.4.
1526 Version 6.1 21-Jun-05
1527 ---------------------
1529 1. There was one reference to the variable "posix" in pcretest.c that was not
1530 surrounded by "#if !defined NOPOSIX".
1532 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1533 the cross-check on the old pcre_info() function, for the benefit of the
1534 cut-down version of PCRE that is currently imported into Exim.
1536 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1537 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1538 patterns that start (?i) but I don't think that matters. The use of (?i) is
1539 just an example; this all applies to the other options as well.
1541 4. Since libtool seems to echo the compile commands it is issuing, the output
1542 from "make" can be reduced a bit by putting "@" in front of each libtool
1543 compile command.
1545 5. Patch from the folks at Google for configure.in to be a bit more thorough
1546 in checking for a suitable C++ installation before trying to compile the
1547 C++ stuff. This should fix a reported problem when a compiler was present,
1548 but no suitable headers.
1550 6. The man pages all had just "PCRE" as their title. I have changed them to
1551 be the relevant file name. I have also arranged that these names are
1552 retained in the file doc/pcre.txt, which is a concatenation in text format
1553 of all the man pages except the little individual ones for each function.
1555 7. The NON-UNIX-USE file had not been updated for the different set of source
1556 files that come with release 6. I also added a few comments about the C++
1557 wrapper.
1560 Version 6.0 07-Jun-05
1561 ---------------------
1563 1. Some minor internal re-organization to help with my DFA experiments.
1565 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1566 didn't matter for the library itself when fully configured, but did matter
1567 when compiling without UCP support, or within Exim, where the ucp files are
1568 not imported.
1570 3. Refactoring of the library code to split up the various functions into
1571 different source modules. The addition of the new DFA matching code (see
1572 below) to a single monolithic source would have made it really too
1573 unwieldy, quite apart from causing all the code to be include in a
1574 statically linked application, when only some functions are used. This is
1575 relevant even without the DFA addition now that patterns can be compiled in
1576 one application and matched in another.
1578 The downside of splitting up is that there have to be some external
1579 functions and data tables that are used internally in different modules of
1580 the library but which are not part of the API. These have all had their
1581 names changed to start with "_pcre_" so that they are unlikely to clash
1582 with other external names.
1584 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1585 a different (DFA) algorithm. Although it is slower than the original
1586 function, it does have some advantages for certain types of matching
1587 problem.
1589 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1590 including restarting after a partial match.
1592 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1593 defined when compiling for Windows was sent to me. I have put it into the
1594 code, though I have no means of testing or verifying it.
1596 7. Added the pcre_refcount() auxiliary function.
1598 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1599 match before or at the first newline in the subject string. In pcretest,
1600 the /f option on a pattern can be used to set this.
1602 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1603 would behave wrongly. This has been present in PCRE since release 4.0.
1605 10. A number of changes to the pcregrep command:
1607 (a) Refactored how -x works; insert ^(...)$ instead of setting
1608 PCRE_ANCHORED and checking the length, in preparation for adding
1609 something similar for -w.
1611 (b) Added the -w (match as a word) option.
1613 (c) Refactored the way lines are read and buffered so as to have more
1614 than one at a time available.
1616 (d) Implemented a pcregrep test script.
1618 (e) Added the -M (multiline match) option. This allows patterns to match
1619 over several lines of the subject. The buffering ensures that at least
1620 8K, or the rest of the document (whichever is the shorter) is available
1621 for matching (and similarly the previous 8K for lookbehind assertions).
1623 (f) Changed the --help output so that it now says
1625 -w, --word-regex(p)
1627 instead of two lines, one with "regex" and the other with "regexp"
1628 because that confused at least one person since the short forms are the
1629 same. (This required a bit of code, as the output is generated
1630 automatically from a table. It wasn't just a text change.)
1632 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1633 option but starts with a hyphen. Could be a pattern or a path name
1634 starting with a hyphen, for instance.
1636 (h) "-" can be given as a file name to represent stdin.
1638 (i) When file names are being printed, "(standard input)" is used for
1639 the standard input, for compatibility with GNU grep. Previously
1640 "<stdin>" was used.
1642 (j) The option --label=xxx can be used to supply a name to be used for
1643 stdin when file names are being printed. There is no short form.
1645 (k) Re-factored the options decoding logic because we are going to add
1646 two more options that take data. Such options can now be given in four
1647 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1649 (l) Added the -A, -B, and -C options for requesting that lines of context
1650 around matches be printed.
1652 (m) Added the -L option to print the names of files that do not contain
1653 any matching lines, that is, the complement of -l.
1655 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1656 continue to scan other files.
1658 (o) The -s option was incorrectly implemented. For compatibility with other
1659 greps, it now suppresses the error message for a non-existent or non-
1660 accessible file (but not the return code). There is a new option called
1661 -q that suppresses the output of matching lines, which was what -s was
1662 previously doing.
1664 (p) Added --include and --exclude options to specify files for inclusion
1665 and exclusion when recursing.
1667 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1668 Hopefully, it now does.
1670 12. Missing cast in pcre_study().
1672 13. Added an "uninstall" target to the makefile.
1674 14. Replaced "extern" in the function prototypes in Makefile.in with
1675 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1676 world, but is set differently for Windows.
1678 15. Added a second compiling function called pcre_compile2(). The only
1679 difference is that it has an extra argument, which is a pointer to an
1680 integer error code. When there is a compile-time failure, this is set
1681 non-zero, in addition to the error test pointer being set to point to an
1682 error message. The new argument may be NULL if no error number is required
1683 (but then you may as well call pcre_compile(), which is now just a
1684 wrapper). This facility is provided because some applications need a
1685 numeric error indication, but it has also enabled me to tidy up the way
1686 compile-time errors are handled in the POSIX wrapper.
1688 16. Added VPATH=.libs to the makefile; this should help when building with one
1689 prefix path and installing with another. (Or so I'm told by someone who
1690 knows more about this stuff than I do.)
1692 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1693 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1694 match everything, including newlines. This is not POSIX-compatible, but
1695 somebody wanted the feature. From pcretest it can be activated by using
1696 both the P and the s flags.
1698 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1700 19. libpcre.pc was being incorrectly installed as executable.
1702 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1703 it now also looks for '\r' so that it will work unmodified on Windows.
1705 21. Added Google's contributed C++ wrapper to the distribution.
1707 22. Added some untidy missing memory free() calls in pcretest, to keep
1708 Electric Fence happy when testing.
1712 Version 5.0 13-Sep-04
1713 ---------------------
1715 1. Internal change: literal characters are no longer packed up into items
1716 containing multiple characters in a single byte-string. Each character
1717 is now matched using a separate opcode. However, there may be more than one
1718 byte in the character in UTF-8 mode.
1720 2. The pcre_callout_block structure has two new fields: pattern_position and
1721 next_item_length. These contain the offset in the pattern to the next match
1722 item, and its length, respectively.
1724 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1725 insertion of callouts before each pattern item. Added the /C option to
1726 pcretest to make use of this.
1728 4. On the advice of a Windows user, the lines
1730 #if defined(_WIN32) || defined(WIN32)
1731 _setmode( _fileno( stdout ), 0x8000 );
1732 #endif /* defined(_WIN32) || defined(WIN32) */
1734 have been added to the source of pcretest. This apparently does useful
1735 magic in relation to line terminators.
1737 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1738 for the benefit of those environments where the "b" makes a difference.
1740 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1741 to know about it. I have put a hack into configure.in that adds in code
1742 to set GCC=yes if CC=icc. This seems to end up at a point in the
1743 generated configure script that is early enough to affect the setting of
1744 compiler options, which is what is needed, but I have no means of testing
1745 whether it really works. (The user who reported this had patched the
1746 generated configure script, which of course I cannot do.)
1748 LATER: After change 22 below (new libtool files), the configure script
1749 seems to know about icc (and also ecc). Therefore, I have commented out
1750 this hack in configure.in.
1752 7. Added support for pkg-config (2 patches were sent in).
1754 8. Negated POSIX character classes that used a combination of internal tables
1755 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1756 [[:^ascii]]. Typically, they would match almost any characters. The other
1757 POSIX classes were not broken in this way.
1759 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1760 to find the match, as PCRE was deluded into thinking that the match had to
1761 start at the start point or following a newline. The same bug applied to
1762 patterns with negative forward assertions or any backward assertions
1763 preceding ".*" at the start, unless the pattern required a fixed first
1764 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1766 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1767 starting at the last subject character, bytes beyond the end of the subject
1768 string were read.
1770 11. Renamed the variable "class" as "classbits" to make life easier for C++
1771 users. (Previously there was a macro definition, but it apparently wasn't
1772 enough.)
1774 12. Added the new field "tables" to the extra data so that tables can be passed
1775 in at exec time, or the internal tables can be re-selected. This allows
1776 a compiled regex to be saved and re-used at a later time by a different
1777 program that might have everything at different addresses.
1779 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1780 -R library as well as a -L library.
1782 14. The debugging options of pcretest (-d on the command line or D on a
1783 pattern) showed incorrect output for anything following an extended class
1784 that contained multibyte characters and which was followed by a quantifier.
1786 15. Added optional support for general category Unicode character properties
1787 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1788 support. It adds about 90K to the size of the library. The meanings of the
1789 inbuilt class escapes such as \d and \s have NOT been changed.
1791 16. Updated pcredemo.c to include calls to free() to release the memory for the
1792 compiled pattern.
1794 17. The generated file chartables.c was being created in the source directory
1795 instead of in the building directory. This caused the build to fail if the
1796 source directory was different from the building directory, and was
1797 read-only.
1799 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1800 file. No doubt somebody will tell me if they don't make sense... Also added
1801 Dan Mooney's comments about building on OpenVMS.
1803 19. Added support for partial matching via the PCRE_PARTIAL option for
1804 pcre_exec() and the \P data escape in pcretest.
1806 20. Extended pcretest with 3 new pattern features:
1808 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1809 write the compiled pattern to the file whose name is "rest-of-line".
1810 This is a straight binary dump of the data, with the saved pointer to
1811 the character tables forced to be NULL. The study data, if any, is
1812 written too. After writing, pcretest reads a new pattern.
1814 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1815 compiled pattern from the given file. There must not be any
1816 occurrences of "<" in the file name (pretty unlikely); if there are,
1817 pcretest will instead treat the initial "<" as a pattern delimiter.
1818 After reading in the pattern, pcretest goes on to read data lines as
1819 usual.
1821 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1822 and 16-bit fields in a compiled pattern, to simulate a pattern that
1823 was compiled on a host of opposite endianness.
1825 21. The pcre-exec() function can now cope with patterns that were compiled on
1826 hosts of opposite endianness, with this restriction:
1828 As for any compiled expression that is saved and used later, the tables
1829 pointer field cannot be preserved; the extra_data field in the arguments
1830 to pcre_exec() should be used to pass in a tables address if a value
1831 other than the default internal tables were used at compile time.
1833 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1834 now diagnosed as an error. Previously, most of the time, a negative number
1835 would have been treated as zero, but if in addition "ovector" was passed as
1836 NULL, a crash could occur.
1838 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1839 new versions from the libtool 1.5 distribution (the last one is a copy of
1840 a file called libtool.m4). This seems to have fixed the need to patch
1841 "configure" to support Darwin 1.3 (which I used to do). However, I still
1842 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1843 workstation).
1845 24. Changed the PCRE licence to be the more standard "BSD" licence.
1848 Version 4.5 01-Dec-03
1849 ---------------------
1851 1. There has been some re-arrangement of the code for the match() function so
1852 that it can be compiled in a version that does not call itself recursively.
1853 Instead, it keeps those local variables that need separate instances for
1854 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1855 needs to "recurse". Keeping track of where control must go is done by means
1856 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1857 hide most of the details from the main code, and operates only if
1858 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1859 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1860 operating.
1862 To make it easier for callers to provide specially tailored get/free
1863 functions for this usage, two new functions, pcre_stack_malloc, and
1864 pcre_stack_free, are used. They are always called in strict stacking order,
1865 and the size of block requested is always the same.
1867 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1868 PCRE has been compiled to use the stack or the heap for recursion. The
1869 -C option of pcretest uses this to show which version is compiled.
1871 A new data escape \S, is added to pcretest; it causes the amounts of store
1872 obtained and freed by both kinds of malloc/free at match time to be added
1873 to the output.
1875 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1876 what's available on my current Linux desktop machine.
1878 3. When matching a UTF-8 string, the test for a valid string at the start has
1879 been extended. If start_offset is not zero, PCRE now checks that it points
1880 to a byte that is the start of a UTF-8 character. If not, it returns
1881 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1882 this is necessary because there may be backward assertions in the pattern.
1883 When matching the same subject several times, it may save resources to use
1884 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1886 4. The code for checking the validity of UTF-8 strings has been tightened so
1887 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1888 containing "overlong sequences".
1890 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1891 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1892 should have been "&", but it just so happened that all the cases this let
1893 through by mistake were picked up later in the function.
1895 6. I had used a variable called "isblank" - this is a C99 function, causing
1896 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1898 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1899 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1900 so the version is shown; (c) stop "make test" echoing "./RunTest".
1902 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1904 9. The prototype for memmove() for systems that don't have it was using
1905 size_t, but the inclusion of the header that defines size_t was later. I've
1906 moved the #includes for the C headers earlier to avoid this.
1908 10. Added some adjustments to the code to make it easier to compiler on certain
1909 special systems:
1911 (a) Some "const" qualifiers were missing.
1912 (b) Added the macro EXPORT before all exported functions; by default this
1913 is defined to be empty.
1914 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1915 that it reads its output file name as an argument instead of writing
1916 to the standard output and assuming this can be redirected.
1918 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1919 class containing characters with values greater than 255, PCRE compilation
1920 went into a loop.
1922 12. A recursive reference to a subpattern that was within another subpattern
1923 that had a minimum quantifier of zero caused PCRE to crash. For example,
1924 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1925 recursion. If the recursively-called subpattern itself had a zero repeat,
1926 that was OK.
1928 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1929 buffer into which it was copied (for escape processing) was still set at
1930 1024, so long lines caused crashes.
1932 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1933 "internal error: code overflow...". This applied to any character class
1934 that was followed by a possessive quantifier.
1936 15. Modified the Makefile to add libpcre.la as a prerequisite for
1937 libpcreposix.la because I was told this is needed for a parallel build to
1938 work.
1940 16. If a pattern that contained .* following optional items at the start was
1941 studied, the wrong optimizing data was generated, leading to matching
1942 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1943 matching string must start with a or b or c. The correct conclusion for
1944 this pattern is that a match can start with any character.
1947 Version 4.4 13-Aug-03
1948 ---------------------
1950 1. In UTF-8 mode, a character class containing characters with values between
1951 127 and 255 was not handled correctly if the compiled pattern was studied.
1952 In fixing this, I have also improved the studying algorithm for such
1953 classes (slightly).
1955 2. Three internal functions had redundant arguments passed to them. Removal
1956 might give a very teeny performance improvement.
1958 3. Documentation bug: the value of the capture_top field in a callout is *one
1959 more than* the number of the hightest numbered captured substring.
1961 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1962 in incorrectly linking with a previously installed version. They now link
1963 explicitly with libpcre.la.
1965 5. configure.in no longer needs to recognize Cygwin specially.
1967 6. A problem in pcre.in for Windows platforms is fixed.
1969 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1970 pcretest, it used to include the size of the study block as part of its
1971 output. Unfortunately, the structure contains a field that has a different
1972 size on different hardware architectures. This meant that the tests that
1973 showed this size failed. As the block is currently always of a fixed size,
1974 this information isn't actually particularly useful in pcretest output, so
1975 I have just removed it.
1977 8. Three pre-processor statements accidentally did not start in column 1.
1978 Sadly, there are *still* compilers around that complain, even though
1979 standard C has not required this for well over a decade. Sigh.
1981 9. In pcretest, the code for checking callouts passed small integers in the
1982 callout_data field, which is a void * field. However, some picky compilers
1983 complained about the casts involved for this on 64-bit systems. Now
1984 pcretest passes the address of the small integer instead, which should get
1985 rid of the warnings.
1987 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1988 both compile and run time, and gives an error if an invalid UTF-8 sequence
1989 is found. There is a option for disabling this check in cases where the
1990 string is known to be correct and/or the maximum performance is wanted.
1992 11. In response to a bug report, I changed one line in Makefile.in from
1994 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1995 to
1996 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1998 to look similar to other lines, but I have no way of telling whether this
1999 is the right thing to do, as I do not use Windows. No doubt I'll get told
2000 if it's wrong...
2003 Version 4.3 21-May-03
2004 ---------------------
2006 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2007 Makefile.
2009 2. Some refactoring to improve the quality of the code:
2011 (i) The utf8_table... variables are now declared "const".
2013 (ii) The code for \cx, which used the "case flipping" table to upper case
2014 lower case letters, now just substracts 32. This is ASCII-specific,
2015 but the whole concept of \cx is ASCII-specific, so it seems
2016 reasonable.
2018 (iii) PCRE was using its character types table to recognize decimal and
2019 hexadecimal digits in the pattern. This is silly, because it handles
2020 only 0-9, a-f, and A-F, but the character types table is locale-
2021 specific, which means strange things might happen. A private
2022 table is now used for this - though it costs 256 bytes, a table is
2023 much faster than multiple explicit tests. Of course, the standard
2024 character types table is still used for matching digits in subject
2025 strings against \d.
2027 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2028 ending in _t are). So I've renamed it as ESC_tee.
2030 3. The first argument for regexec() in the POSIX wrapper should have been
2031 defined as "const".
2033 4. Changed pcretest to use malloc() for its buffers so that they can be
2034 Electric Fenced for debugging.
2036 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2037 to read one or more bytes before the start of the subject string. Often this
2038 had no effect on PCRE's behaviour, but in some circumstances it could
2039 provoke a segmentation fault.
2041 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2042 to try to read one or more bytes before the start of the subject string.
2044 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2045 UTF-8 support could misbehave in various ways if the subject string
2046 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2047 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2048 back over UTF-8 characters.)
2051 Version 4.2 14-Apr-03
2052 ---------------------
2054 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2056 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2057 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2058 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2059 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2060 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2062 Note: automatic setting of the BUILD variables is not yet working
2063 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2064 compile-time but not at link-time
2065 [LINK]: use for linking executables only
2066 make different versions for Windows and non-Windows
2067 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2068 libraries
2069 [LINK_FOR_BUILD]: new variable
2070 [OBJEXT]: use throughout
2071 [EXEEXT]: use throughout
2072 <winshared>: new target
2073 <wininstall>: new target
2074 <dftables.o>: use native compiler
2075 <dftables>: use native linker
2076 <install>: handle Windows platform correctly
2077 <clean>: ditto
2078 <check>: ditto
2079 copy DLL to top builddir before testing
2081 As part of these changes, -no-undefined was removed again. This was reported
2082 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2083 in any case.
2085 3. Some tidies to get rid of compiler warnings:
2087 . In the match_data structure, match_limit was an unsigned long int, whereas
2088 match_call_count was an int. I've made them both unsigned long ints.
2090 . In pcretest the fact that a const uschar * doesn't automatically cast to
2091 a void * provoked a warning.
2093 . Turning on some more compiler warnings threw up some "shadow" variables
2094 and a few more missing casts.
2096 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2097 option, a class that contained a single character with a value between 128
2098 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2100 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2101 option, a class that contained several characters, but with at least one
2102 whose value was between 128 and 255 caused PCRE to crash.
2105 Version 4.1 12-Mar-03
2106 ---------------------
2108 1. Compiling with gcc -pedantic found a couple of places where casts were
2109 needed, and a string in dftables.c that was longer than standard compilers are
2110 required to support.
2112 2. Compiling with Sun's compiler found a few more places where the code could
2113 be tidied up in order to avoid warnings.
2115 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2116 first of these names is deprecated in the latest Autoconf in favour of the name
2117 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2118 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2119 analogy I have changed it to CFLAGS_FOR_BUILD.
2121 4. Added -no-undefined to the linking command in the Makefile, because this is
2122 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2123 linking step for the pcreposix library.
2125 5. PCRE was failing to diagnose the case of two named groups with the same
2126 name.
2128 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2129 literal character that is needed in the subject for a match, and scans along to
2130 ensure that it is present before embarking on the full matching process. This
2131 saves time in cases of nested unlimited repeats that are never going to match.
2132 Problem: the scan can take a lot of time if the subject is very long (e.g.
2133 megabytes), thus penalizing straightforward matches. It is now done only if the
2134 amount of subject to be scanned is less than 1000 bytes.
2136 7. A lesser problem with the same optimization is that it was recording the
2137 first character of an anchored pattern as "needed", thus provoking a search
2138 right along the subject, even when the first match of the pattern was going to
2139 fail. The "needed" character is now not set for anchored patterns, unless it
2140 follows something in the pattern that is of non-fixed length. Thus, it still
2141 fulfils its original purpose of finding quick non-matches in cases of nested
2142 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2145 Version 4.0 17-Feb-03
2146 ---------------------
2148 1. If a comment in an extended regex that started immediately after a meta-item
2149 extended to the end of string, PCRE compiled incorrect data. This could lead to
2150 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2152 2. Moved to autoconf 2.53 and libtool 1.4.2.
2154 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2155 the special perltest8 script is no longer needed - all the tests can be run
2156 from a single perltest script.
2158 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2159 by \s. It has now been removed in PCRE. This means it isn't recognized as
2160 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2161 class [:space:] *does* include VT, thereby creating a mess.
2163 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2164 space and tab.
2166 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2167 its new features into the main test script, reducing the number of scripts.
2169 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2170 were backward compatible, and made the (?i) apply to the whole pattern, as if
2171 /i were given. Now it behaves more logically, and applies the option setting
2172 only to what follows. PCRE has been changed to follow suit. However, if it
2173 finds options settings right at the start of the pattern, it extracts them into
2174 the global options, as before. Thus, they show up in the info data.
2176 8. Added support for the \Q...\E escape sequence. Characters in between are
2177 treated as literals. This is slightly different from Perl in that $ and @ are
2178 also handled as literals inside the quotes. In Perl, they will cause variable
2179 interpolation. Note the following examples:
2181 Pattern PCRE matches Perl matches
2183 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2184 \Qabc\$xyz\E abc\$xyz abc\$xyz
2185 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2187 For compatibility with Perl, \Q...\E sequences are recognized inside character
2188 classes as well as outside them.
2190 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2191 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2192 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2193 signed/unsigned warnings.
2195 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2196 option for pcretest, so I've replaced it by a simple function that does just
2197 that job.
2199 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2200 "pcregrep -".
2202 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2203 Java package. This provides some syntactic sugar for simple cases of what my
2204 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2205 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2206 item, you can use this simplified notation. Note that only makes sense with
2207 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2208 greediness, whatever the setting of the PCRE_UNGREEDY option.
2210 13. A change of greediness default within a pattern was not taking effect at
2211 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2212 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2213 was abstracted outside.
2215 14. PCRE now supports the \G assertion. It is true when the current matching
2216 position is at the start point of the match. This differs from \A when the
2217 starting offset is non-zero. Used with the /g option of pcretest (or similar
2218 code), it works in the same way as it does for Perl's /g option. If all
2219 alternatives of a regex begin with \G, the expression is anchored to the start
2220 match position, and the "anchored" flag is set in the compiled expression.
2222 15. Some bugs concerning the handling of certain option changes within patterns
2223 have been fixed. These applied to options other than (?ims). For example,
2224 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2225 been the other way round. Some of this was related to change 7 above.
2227 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2228 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2229 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2230 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2232 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2233 mode. Unlike ".", it always matches newline, whatever the setting of
2234 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2235 assertions. Perl allows it, but it doesn't (in general) work because it can't
2236 calculate the length of the lookbehind. At least, that's the case for Perl
2237 5.8.0 - I've been told they are going to document that it doesn't work in
2238 future.
2240 18. Added an error diagnosis for escapes that PCRE does not support: these are
2241 \L, \l, \N, \P, \p, \U, \u, and \X.
2243 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2244 reading past the end of the pattern in cases such as /[abcd/.
2246 20. PCRE was getting more memory than necessary for patterns with classes that
2247 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2249 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2250 compiling PCRE for use with Virtual Pascal.
2252 22. Small fix to the Makefile to make it work properly if the build is done
2253 outside the source tree.
2255 23. Added a new extension: a condition to go with recursion. If a conditional
2256 subpattern starts with (?(R) the "true" branch is used if recursion has
2257 happened, whereas the "false" branch is used only at the top level.
2259 24. When there was a very long string of literal characters (over 255 bytes
2260 without UTF support, over 250 bytes with UTF support), the computation of how
2261 much memory was required could be incorrect, leading to segfaults or other
2262 strange effects.
2264 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2265 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2266 there was a subsequent back reference to those brackets. This meant that, for
2267 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2268 possible to check for precisely this case. All we can do is abandon the
2269 optimization if .* occurs inside capturing brackets when there are any back
2270 references whatsoever. (See below for a better fix that came later.)
2272 26. The handling of the optimization for finding the first character of a
2273 non-anchored pattern, and for finding a character that is required later in the
2274 match were failing in some cases. This didn't break the matching; it just
2275 failed to optimize when it could. The way this is done has been re-implemented.
2277 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2279 28. Added a new feature that provides some of the functionality that Perl
2280 provides with (?{...}). The facility is termed a "callout". The way it is done
2281 in PCRE is for the caller to provide an optional function, by setting
2282 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2283 global variable. By default it is unset, which disables all calling out. To get
2284 the function called, the regex must include (?C) at appropriate points. This
2285 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2286 This provides a means of identifying different callout points. When PCRE
2287 reaches such a point in the regex, if pcre_callout has been set, the external
2288 function is called. It is provided with data in a structure called
2289 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2290 matching continues; if it returns a non-zero value, the match at the current
2291 point fails. However, backtracking will occur if possible. [This was changed
2292 later and other features added - see item 49 below.]
2294 29. pcretest is upgraded to test the callout functionality. It provides a
2295 callout function that displays information. By default, it shows the start of
2296 the match and the current position in the text. There are some new data escapes
2297 to vary what happens:
2299 \C+ in addition, show current contents of captured substrings
2300 \C- do not supply a callout function
2301 \C!n return 1 when callout number n is reached
2302 \C!n!m return 1 when callout number n is reached for the mth time
2304 30. If pcregrep was called with the -l option and just a single file name, it
2305 output "<stdin>" if a match was found, instead of the file name.
2307 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2308 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2309 pcre_exec(). This saves a malloc/free per call. The default value of
2310 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2311 when configuring.
2313 32. The default maximum size of a compiled pattern is 64K. There have been a
2314 few cases of people hitting this limit. The code now uses macros to handle the
2315 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2316 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2317 configuring. Tests 2 and 5 work only with 2-byte links because they output
2318 debugging information about compiled patterns.
2320 33. Internal code re-arrangements:
2322 (a) Moved the debugging function for printing out a compiled regex into
2323 its own source file (printint.c) and used #include to pull it into
2324 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2325 separate copies.
2327 (b) Defined the list of op-code names for debugging as a macro in
2328 internal.h so that it is next to the definition of the opcodes.
2330 (c) Defined a table of op-code lengths for simpler skipping along compiled
2331 code. This is again a macro in internal.h so that it is next to the
2332 definition of the opcodes.
2334 34. Added support for recursive calls to individual subpatterns, along the
2335 lines of Robin Houston's patch (but implemented somewhat differently).
2337 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2338 allow it to read and process whole directories in Win32. This code was
2339 contributed by Lionel Fourquaux; it has not been tested by me.
2341 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2342 used to name a group. Names consist of alphanumerics and underscores, and must
2343 be unique. Back references use the syntax (?P=name) and recursive calls use
2344 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2345 numbers. The function pcre_fullinfo() can be used after compilation to extract
2346 a name/number map. There are three relevant calls:
2348 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2349 PCRE_INFO_NAMECOUNT yields the number of entries
2350 PCRE_INFO_NAMETABLE yields a pointer to the map.
2352 The map is a vector of fixed-size entries. The size of each entry depends on
2353 the length of the longest name used. The first two bytes of each entry are the
2354 group number, most significant byte first. There follows the corresponding
2355 name, zero terminated. The names are in alphabetical order.
2357 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2358 case instead of 255. Making it the same both with and without UTF-8 support
2359 means that the same test output works with both.
2361 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2362 calling malloc() with a zero argument.
2364 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2365 optimization. I've improved things by keeping a bitmap of backreferences with
2366 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2367 fact referenced, the optimization can be applied. It is unlikely that a
2368 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2369 the match to follow \n) will appear inside brackets with a number greater than
2370 31, but if it does, any back reference > 31 suppresses the optimization.
2372 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2373 of disabling numbered capturing parentheses. Any opening parenthesis that is
2374 not followed by ? behaves as if it were followed by ?: but named parentheses
2375 can still be used for capturing (and they will acquire numbers in the usual
2376 way).
2378 41. Redesigned the return codes from the match() function into yes/no/error so
2379 that errors can be passed back from deep inside the nested calls. A malloc
2380 failure while inside a recursive subpattern call now causes the
2381 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2383 42. It is now possible to set a limit on the number of times the match()
2384 function is called in a call to pcre_exec(). This facility makes it possible to
2385 limit the amount of recursion and backtracking, though not in a directly
2386 obvious way, because the match() function is used in a number of different
2387 circumstances. The count starts from zero for each position in the subject
2388 string (for non-anchored patterns). The default limit is, for compatibility, a
2389 large number, namely 10 000 000. You can change this in two ways:
2391 (a) When configuring PCRE before making, you can use --with-match-limit=n
2392 to set a default value for the compiled library.
2394 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2395 a different value is set. See 45 below.
2397 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2399 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2400 of things that can be changed at compile time. The first argument specifies
2401 what is wanted and the second points to where the information is to be placed.
2402 The current list of available information is:
2406 The output is an integer that is set to one if UTF-8 support is available;
2407 otherwise it is set to zero.
2411 The output is an integer that it set to the value of the code that is used for
2412 newline. It is either LF (10) or CR (13).
2416 The output is an integer that contains the number of bytes used for internal
2417 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2421 The output is an integer that contains the threshold above which the POSIX
2422 interface uses malloc() for output vectors. See item 31 above.
2426 The output is an unsigned integer that contains the default limit of the number
2427 of match() calls in a pcre_exec() execution. See 42 above.
2429 44. pcretest has been upgraded by the addition of the -C option. This causes it
2430 to extract all the available output from the new pcre_config() function, and to
2431 output it. The program then exits immediately.
2433 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2434 order to support additional features. One way would have been to define
2435 pcre_exec2() (for example) with extra arguments, but this would not have been
2436 extensible, and would also have required all calls to the original function to
2437 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2438 is used for passing in "extra" data from pcre_study().
2440 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2441 contains the following fields:
2443 flags a bitmap indicating which of the following fields are set
2444 study_data opaque data from pcre_study()
2445 match_limit a way of specifying a limit on match() calls for a specific
2446 call to pcre_exec()
2447 callout_data data for callouts (see 49 below)
2449 The flag bits are also defined in pcre.h, and are
2455 The pcre_study() function now returns one of these new pcre_extra blocks, with
2456 the actual study data pointed to by the study_data field, and the
2457 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2458 before. That is, this change is entirely upwards-compatible and requires no
2459 change to existing code.
2461 If you want to pass in additional data to pcre_exec(), you can either place it
2462 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2463 block.
2465 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2466 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2467 times with different match limits, until it finds the minimum value needed for
2468 pcre_exec() to complete. The value is then output. This can be instructive; for
2469 most simple matches the number is quite small, but for pathological cases it
2470 gets very large very quickly.
2472 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2473 returns the size of the data block pointed to by the study_data field in a
2474 pcre_extra block, that is, the value that was passed as the argument to
2475 pcre_malloc() when PCRE was getting memory in which to place the information
2476 created by pcre_study(). The fourth argument should point to a size_t variable.
2477 pcretest has been extended so that this information is shown after a successful
2478 pcre_study() call when information about the compiled regex is being displayed.
2480 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2481 because what follows is always an absolute path. (Later: it turns out that this
2482 is more than cosmetic for MinGW, because it doesn't like empty path
2483 components.)
2485 49. Some changes have been made to the callout feature (see 28 above):
2487 (i) A callout function now has three choices for what it returns:
2489 0 => success, carry on matching
2490 > 0 => failure at this point, but backtrack if possible
2491 < 0 => serious error, return this value from pcre_exec()
2493 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2494 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2495 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2496 use by callout functions. It will never be used by PCRE itself.
2498 (ii) The pcre_extra structure (see 45 above) has a void * field called
2499 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2500 pcre_callout_block structure has a field of the same name. The contents of
2501 the field passed in the pcre_extra structure are passed to the callout
2502 function in the corresponding field in the callout block. This makes it
2503 easier to use the same callout-containing regex from multiple threads. For
2504 testing, the pcretest program has a new data escape
2506 \C*n pass the number n (may be negative) as callout_data
2508 If the callout function in pcretest receives a non-zero value as
2509 callout_data, it returns that value.
2511 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2512 there were some redundant $(CFLAGS) in commands that are now specified as
2513 $(LINK), which already includes $(CFLAGS).
2515 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2516 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2517 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2518 one-byte characters throughout. Note that case-insensitive matching applies
2519 only to characters whose values are less than 256. PCRE doesn't support the
2520 notion of cases for higher-valued characters.
2522 (i) A character class whose characters are all within 0-255 is handled as
2523 a bit map, and the map is inverted for negative classes. Previously, a
2524 character > 255 always failed to match such a class; however it should
2525 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2527 (ii) A negated character class with a single character < 255 is coded as
2528 "not this character" (OP_NOT). This wasn't working properly when the test
2529 character was multibyte, either singly or repeated.
2531 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2532 mode, for example: \x{100}{2,3}.
2534 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2535 singly or repeated) now correctly test multibyte characters. However,
2536 PCRE doesn't recognize any characters with values greater than 255 as
2537 digits, spaces, or word characters. Such characters always match \D, \S,
2538 and \W, and never match \d, \s, or \w.
2540 (v) Classes may now contain characters and character ranges with values
2541 greater than 255. For example: [ab\x{100}-\x{400}].
2543 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2544 PCRE in UTF-8 mode.
2546 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2547 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2548 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2549 value.)
2551 53. The single man page has become too large. I have therefore split it up into
2552 a number of separate man pages. These also give rise to individual HTML pages;
2553 these are now put in a separate directory, and there is an index.html page that
2554 lists them all. Some hyperlinking between the pages has been installed.
2556 54. Added convenience functions for handling named capturing parentheses.
2558 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2559 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2560 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2561 are faulted.
2563 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2564 calling configure. These values are used when compiling the dftables.c program
2565 which is run to generate the source of the default character tables. They
2566 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2567 you will need to set these values.
2569 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2572 Version 3.9 02-Jan-02
2573 ---------------------
2575 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2577 2. If --disable-static was given, the building process failed when trying to
2578 build pcretest and pcregrep. (For some reason it was using libtool to compile
2579 them, which is not right, as they aren't part of the library.)
2582 Version 3.8 18-Dec-01
2583 ---------------------
2585 1. The experimental UTF-8 code was completely screwed up. It was packing the
2586 bytes in the wrong order. How dumb can you get?
2589 Version 3.7 29-Oct-01
2590 ---------------------
2592 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2593 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2594 this didn't happen under Solaris 8, where I normally test things.
2596 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2597 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2598 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2599 this fix an hour or so after the initial 3.7 release.)
2602 Version 3.6 23-Oct-01
2603 ---------------------
2605 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2606 offsets passed as NULL with zero offset count.
2608 2. The config.guess and config.sub files had not been updated when I moved to
2609 the latest autoconf.
2612 Version 3.5 15-Aug-01
2613 ---------------------
2615 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2616 had been forgotten.
2618 2. By using declared but undefined structures, we can avoid using "void"
2619 definitions in pcre.h while keeping the internal definitions of the structures
2620 private.
2622 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2623 user point of view, this means that both static and shared libraries are built
2624 by default, but this can be individually controlled. More of the work of
2625 handling this static/shared cases is now inside libtool instead of PCRE's make
2626 file.
2628 4. The pcretest utility is now installed along with pcregrep because it is
2629 useful for users (to test regexs) and by doing this, it automatically gets
2630 relinked by libtool. The documentation has been turned into a man page, so
2631 there are now .1, .txt, and .html versions in /doc.
2633 5. Upgrades to pcregrep:
2634 (i) Added long-form option names like gnu grep.
2635 (ii) Added --help to list all options with an explanatory phrase.
2636 (iii) Added -r, --recursive to recurse into sub-directories.
2637 (iv) Added -f, --file to read patterns from a file.
2639 6. pcre_exec() was referring to its "code" argument before testing that
2640 argument for NULL (and giving an error if it was NULL).
2642 7. Upgraded Makefile.in to allow for compiling in a different directory from
2643 the source directory.
2645 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2646 options bits, the pointer it was passed was to an int instead of to an unsigned
2647 long int. This mattered only on 64-bit systems.
2649 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2650 generated) instead of pcre.in, which it its source. Also made the same change
2651 in several of the .c files.
2653 10. A new release of gcc defines printf() as a macro, which broke pcretest
2654 because it had an ifdef in the middle of a string argument for printf(). Fixed
2655 by using separate calls to printf().
2657 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2658 script, to force use of CR or LF instead of \n in the source. On non-Unix
2659 systems, the value can be set in config.h.
2661 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2662 absolute limit. Changed the text of the error message to make this clear, and
2663 likewise updated the man page.
2665 13. The limit of 99 on the number of capturing subpatterns has been removed.
2666 The new limit is 65535, which I hope will not be a "real" limit.
2669 Version 3.4 22-Aug-00
2670 ---------------------
2672 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2674 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2677 Version 3.3 01-Aug-00
2678 ---------------------
2680 1. If an octal character was given, but the value was greater than \377, it
2681 was not getting masked to the least significant bits, as documented. This could
2682 lead to crashes in some systems.
2684 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2685 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2687 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2688 These just pass their arguments on to (pcre_free)(), but they are provided
2689 because some uses of PCRE bind it to non-C systems that can call its functions,
2690 but cannot call free() or pcre_free() directly.
2692 4. Add "make test" as a synonym for "make check". Corrected some comments in
2693 the Makefile.
2695 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2696 Makefile.
2698 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2699 command called pgrep for grepping around the active processes.
2701 7. Added the beginnings of support for UTF-8 character strings.
2703 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2704 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2705 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2706 out for the ar command.)
2709 Version 3.2 12-May-00
2710 ---------------------
2712 This is purely a bug fixing release.
2714 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2715 of ZA. This was just one example of several cases that could provoke this bug,
2716 which was introduced by change 9 of version 2.00. The code for breaking
2717 infinite loops after an iteration that matches an empty string was't working
2718 correctly.
2720 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2721 when matched against abbab (for example). After matching an empty string, it
2722 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2723 caused it to match further down the string than it should.
2725 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2726 was there because it doesn't seem to be needed, and it causes trouble on some
2727 systems, as it is not a Standard C header. It has been removed.
2729 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2730 were reported on the Macintosh. The changes were from
2732 while ((c = *(++ptr)) != 0 && c != '\n');
2733 to
2734 while ((c = *(++ptr)) != 0 && c != '\n') ;
2736 Totally extraordinary, but if that's what it takes...
2738 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2739 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2740 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2741 assumes the way PCRE uses memmove() (always moving upwards).
2743 6. PCRE is being used in one environment where strchr() is not available. There
2744 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2745 faster code anyway.
2748 Version 3.1 09-Feb-00
2749 ---------------------
2751 The only change in this release is the fixing of some bugs in Makefile.in for
2752 the "install" target:
2754 (1) It was failing to install pcreposix.h.
2756 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2759 Version 3.0 01-Feb-00
2760 ---------------------
2762 1. Add support for the /+ modifier to perltest (to output $` like it does in
2763 pcretest).
2765 2. Add support for the /g modifier to perltest.
2767 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2768 matches null strings.
2770 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2771 pattern. Perl treats empty patterns specially - it reuses the most recent
2772 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2773 effect.
2775 5. The POSIX interface was broken in that it was just handing over the POSIX
2776 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2777 required a bigger vector, with some working space on the end. This means that
2778 the POSIX wrapper now has to get and free some memory, and copy the results.
2780 6. Added some simple autoconf support, placing the test data and the
2781 documentation in separate directories, re-organizing some of the
2782 information files, and making it build pcre-config (a GNU standard). Also added
2783 libtool support for building PCRE as a shared library, which is now the
2784 default.
2786 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2787 09 are not valid octal constants. Single digits will be used for minor values
2788 less than 10.
2790 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2791 existing programs that set these in the POSIX interface can use PCRE without
2792 modification.
2794 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2795 return all that pcre_info() returns, plus additional data. The pcre_info()
2796 function is retained for compatibility, but is considered to be obsolete.
2798 10. Added experimental recursion feature (?R) to handle one common case that
2799 Perl 5.6 will be able to do with (?p{...}).
2801 11. Added support for POSIX character classes like [:alpha:], which Perl is
2802 adopting.
2805 Version 2.08 31-Aug-99
2806 ----------------------
2808 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2809 trying to match at the startoffset position, but instead was moving forward to
2810 the next newline as if a previous match had failed.
2812 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2813 and could get into a loop if a null string was matched other than at the start
2814 of the subject.
2816 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2817 be distinguished at compile time, and for completeness also added PCRE_DATE.
2819 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2820 in GnuWin32 environments.
2823 Version 2.07 29-Jul-99
2824 ----------------------
2826 1. The documentation is now supplied in plain text form and HTML as well as in
2827 the form of man page sources.
2829 2. C++ compilers don't like assigning (void *) values to other pointer types.
2830 In particular this affects malloc(). Although there is no problem in Standard
2831 C, I've put in casts to keep C++ compilers happy.
2833 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2834 should be (const char *).
2836 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2837 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2838 However, I haven't made this a standard facility. The documentation doesn't
2839 mention it, and the Makefile doesn't support it.
2841 5. The Makefile now contains an "install" target, with editable destinations at
2842 the top of the file. The pcretest program is not installed.
2844 6. pgrep -V now gives the PCRE version number and date.
2846 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2847 causing the entire string to be ignored, instead of just the last character.
2849 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2850 non-matching string, it can take a very, very long time, even for strings of
2851 quite modest length, because of the nested recursion. PCRE now does better in
2852 some of these cases. It does this by remembering the last required literal
2853 character in the pattern, and pre-searching the subject to ensure it is present
2854 before running the real match. In other words, it applies a heuristic to detect
2855 some types of certain failure quickly, and in the above example, if presented
2856 with a string that has no trailing " it gives "no match" very quickly.
2858 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2859 other alternatives are tried instead.
2862 Version 2.06 09-Jun-99
2863 ----------------------
2865 1. Change pcretest's output for amount of store used to show just the code
2866 space, because the remainder (the data block) varies in size between 32-bit and
2867 64-bit systems.
2869 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2870 start matching at. This allows lookbehinds to work when searching for multiple
2871 occurrences in a string.
2873 3. Added additional options to pcretest for testing multiple occurrences:
2875 /+ outputs the rest of the string that follows a match
2876 /g loops for multiple occurrences, using the new startoffset argument
2877 /G loops for multiple occurrences by passing an incremented pointer
2879 4. PCRE wasn't doing the "first character" optimization for patterns starting
2880 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2881 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2882 the letter 'x'. On long subject strings, this gives a significant speed-up.
2885 Version 2.05 21-Apr-99
2886 ----------------------
2888 1. Changed the type of magic_number from int to long int so that it works
2889 properly on 16-bit systems.
2891 2. Fixed a bug which caused patterns starting with .* not to work correctly
2892 when the subject string contained newline characters. PCRE was assuming
2893 anchoring for such patterns in all cases, which is not correct because .* will
2894 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2895 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2896 must be retried after every newline in the subject.
2899 Version 2.04 18-Feb-99
2900 ----------------------
2902 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2903 computation of the store needed to hold the pattern was incorrect (too large).
2904 If such patterns were nested a few deep, this could multiply and become a real
2905 problem.
2907 2. Added /M option to pcretest to show the memory requirement of a specific
2908 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2910 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2911 compiled in such a way that the backtracking after subsequent failure was
2912 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2913 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2916 Version 2.03 02-Feb-99
2917 ----------------------
2919 1. Fixed typo and small mistake in man page.
2921 2. Added 4th condition (GPL supersedes if conflict) and created separate
2922 LICENCE file containing the conditions.
2924 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2925 Perl, that is the internal \ allows the delimiter to be included in the
2926 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2927 the final delimiter, add \ to the end of the pattern (to test the error).
2929 4. Added the convenience functions for extracting substrings after a successful
2930 match. Updated pcretest to make it able to test these functions.
2933 Version 2.02 14-Jan-99
2934 ----------------------
2936 1. Initialized the working variables associated with each extraction so that
2937 their saving and restoring doesn't refer to uninitialized store.
2939 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2940 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2941 fix the problem.
2943 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2944 calls, and wasn't printing the correct value for compiling calls. Increased the
2945 default value of LOOPREPEAT, and the number of significant figures in the
2946 times.
2948 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2950 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2951 a building problem on Windows NT with a FAT file system.
2954 Version 2.01 21-Oct-98
2955 ----------------------
2957 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2958 to character tables built by pcre_maketables() in the current locale. If NULL
2959 is passed, the default tables are used.
2962 Version 2.00 24-Sep-98
2963 ----------------------
2965 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2966 it any more.
2968 2. Allow quantification of (?>) groups, and make it work correctly.
2970 3. The first character computation wasn't working for (?>) groups.
2972 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2973 end of the subject) and add 5.005's \z, which really does match only at the
2974 very end of the subject.
2976 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2978 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2979 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2980 localized options. All options to pcre_study() were also removed.
2982 7. Add other new features from 5.005:
2984 $(?<= positive lookbehind
2985 $(?<! negative lookbehind
2986 (?imsx-imsx) added the unsetting capability
2987 such a setting is global if at outer level; local otherwise
2988 (?imsx-imsx:) non-capturing groups with option setting
2989 (?(cond)re|re) conditional pattern matching
2991 A backreference to itself in a repeated group matches the previous
2992 captured string.
2994 8. General tidying up of studying (both automatic and via "study")
2995 consequential on the addition of new assertions.
2997 9. As in 5.005, unlimited repeated groups that could match an empty substring
2998 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2999 runtime if any iteration does actually match an empty substring.
3001 10. Include the RunTest script in the distribution.
3003 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3004 discrepancies, some of which were old and were also with respect to 5.004. They
3005 have now been fixed.
3008 Version 1.09 28-Apr-98
3009 ----------------------
3011 1. A negated single character class followed by a quantifier with a minimum
3012 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3013 program crashes, or just wrong answers. This did not apply to negated classes
3014 containing more than one character, or to minima other than one.
3017 Version 1.08 27-Mar-98
3018 ----------------------
3020 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3022 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3023 latter must appear before anything that relies on it in the pattern.
3026 Version 1.07 16-Feb-98
3027 ----------------------
3029 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3030 repeat of a potentially empty string).
3033 Version 1.06 23-Jan-98
3034 ----------------------
3036 1. Added Markus Oberhumer's little patches for C++.
3038 2. Literal strings longer than 255 characters were broken.
3041 Version 1.05 23-Dec-97
3042 ----------------------
3044 1. Negated character classes containing more than one character were failing if
3045 PCRE_CASELESS was set at run time.
3048 Version 1.04 19-Dec-97
3049 ----------------------
3051 1. Corrected the man page, where some "const" qualifiers had been omitted.
3053 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3054 input syntax.
3056 3. Fixed memory leak which occurred when a regex with back references was
3057 matched with an offsets vector that wasn't big enough. The temporary memory
3058 that is used in this case wasn't being freed if the match failed.
3060 4. Tidied pcretest to ensure it frees memory that it gets.
3062 5. Temporary memory was being obtained in the case where the passed offsets
3063 vector was exactly big enough.
3065 6. Corrected definition of offsetof() from change 5 below.
3067 7. I had screwed up change 6 below and broken the rules for the use of
3068 setjmp(). Now fixed.
3071 Version 1.03 18-Dec-97
3072 ----------------------
3074 1. A erroneous regex with a missing opening parenthesis was correctly
3075 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3076 on some systems.
3078 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3079 it was reported that one broken compiler failed on the former because "code" is
3080 also an independent variable.
3082 3. The erroneous regex a[]b caused an array overrun reference.
3084 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3085 fail on data ending with that character. (It was going on too far, and checking
3086 the next character, typically a binary zero.) This was specific to the
3087 optimized code for single-character negative classes.
3089 5. Added a contributed patch from the TIN world which does the following:
3091 + Add an undef for memmove, in case the the system defines a macro for it.
3093 + Add a definition of offsetof(), in case there isn't one. (I don't know
3094 the reason behind this - offsetof() is part of the ANSI standard - but
3095 it does no harm).
3097 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3098 most of the places where whitespace preceded '#'. I have given up and
3099 allowed the remaining 2 cases to be at the margin.
3101 + Rename some variables in pcre to eliminate shadowing. This seems very
3102 pedantic, but does no harm, of course.
3104 6. Moved the call to setjmp() into its own function, to get rid of warnings
3105 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3107 7. Constructs such as \d{8,} were compiling into the equivalent of
3108 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3109 outcome, but in this particular case used more store than had been allocated,
3110 which caused the bug to be discovered because it threw up an internal error.
3112 8. The debugging code in both pcre and pcretest for outputting the compiled
3113 form of a regex was going wrong in the case of back references followed by
3114 curly-bracketed repeats.
3117 Version 1.02 12-Dec-97
3118 ----------------------
3120 1. Typos in pcre.3 and comments in the source fixed.
3122 2. Applied a contributed patch to get rid of places where it used to remove
3123 'const' from variables, and fixed some signed/unsigned and uninitialized
3124 variable warnings.
3126 3. Added the "runtest" target to Makefile.
3128 4. Set default compiler flag to -O2 rather than just -O.
3131 Version 1.01 19-Nov-97
3132 ----------------------
3134 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3135 like /([ab]*)*/, that is, for classes with more than one character in them.
3137 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3138 as /((?>a*))*/ (a PCRE_EXTRA facility).
3141 Version 1.00 18-Nov-97
3142 ----------------------
3144 1. Added compile-time macros to support systems such as SunOS4 which don't have
3145 memmove() or strerror() but have other things that can be used instead.
3147 2. Arranged that "make clean" removes the executables.
3150 Version 0.99 27-Oct-97
3151 ----------------------
3153 1. Fixed bug in code for optimizing classes with only one character. It was
3154 initializing a 32-byte map regardless, which could cause it to run off the end
3155 of the memory it had got.
3157 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3160 Version 0.98 22-Oct-97
3161 ----------------------
3163 1. Fixed bug in code for handling temporary memory usage when there are more
3164 back references than supplied space in the ovector. This could cause segfaults.
3167 Version 0.97 21-Oct-97
3168 ----------------------
3170 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3172 2. Optimized negated single characters not to use a bit map.
3174 3. Brought error texts together as macro definitions; clarified some of them;
3175 fixed one that was wrong - it said "range out of order" when it meant "invalid
3176 escape sequence".
3178 4. Changed some char * arguments to const char *.
3180 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3182 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3183 pcretest.
3186 Version 0.96 16-Oct-97
3187 ----------------------
3189 1. Added a simple "pgrep" utility to the distribution.
3191 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3192 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3193 where "ddd" means "one or more decimal digits".
3195 3. Fixed serious bug. If a pattern had a back reference, but the call to
3196 pcre_exec() didn't supply a large enough ovector to record the related
3197 identifying subpattern, the match always failed. PCRE now remembers the number
3198 of the largest back reference, and gets some temporary memory in which to save
3199 the offsets during matching if necessary, in order to ensure that
3200 backreferences always work.
3202 4. Increased the compatibility with Perl in a number of ways:
3204 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3205 to request this handling. The option can be set at compile or exec time.
3207 (b) $ matches before a terminating newline by default; an option
3208 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3209 mode). The option can be set at compile or exec time.
3211 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3212 the same as Perl's. If the decimal number it represents is less than 10
3213 or there aren't that many previous left capturing parentheses, an octal
3214 escape is read. Inside a character class, it's always an octal escape,
3215 even if it is a single digit.
3217 (d) An escaped but undefined alphabetic character is taken as a literal,
3218 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3219 escapes.
3221 (e) {0} is now permitted. (The previous item is removed from the compiled
3222 pattern).
3224 5. Changed all the names of code files so that the basic parts are no longer
3225 than 10 characters, and abolished the teeny "globals.c" file.
3227 6. Changed the handling of character classes; they are now done with a 32-byte
3228 bit map always.
3230 7. Added the -d and /D options to pcretest to make it possible to look at the
3231 internals of compilation without having to recompile pcre.
3234 Version 0.95 23-Sep-97
3235 ----------------------
3237 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3238 \x20 at the start of a run of normal characters. These were being treated as
3239 real characters, instead of the source characters being re-checked.
3242 Version 0.94 18-Sep-97
3243 ----------------------
3245 1. The functions are now thread-safe, with the caveat that the global variables
3246 containing pointers to malloc() and free() or alternative functions are the
3247 same for all threads.
3249 2. Get pcre_study() to generate a bitmap of initial characters for non-
3250 anchored patterns when this is possible, and use it if passed to pcre_exec().
3253 Version 0.93 15-Sep-97
3254 ----------------------
3256 1. /(b)|(:+)/ was computing an incorrect first character.
3258 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3259 but not actually doing anything yet.
3261 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3262 as Perl does (e.g. [-az] or [az-]).
3264 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3265 all possible positions.
3267 5. Split up into different modules to avoid including unneeded functions in a
3268 compiled binary. However, compile and exec are still in one module. The "study"
3269 function is split off.
3271 6. The character tables are now in a separate module whose source is generated
3272 by an auxiliary program - but can then be edited by hand if required. There are
3273 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3274 toupper() in the code.
3276 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3277 make them global. Abolish the function for setting them, as the caller can now
3278 set them directly.
3281 Version 0.92 11-Sep-97
3282 ----------------------
3284 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3285 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3287 2. Caseless matching was not working in character classes if the characters in
3288 the pattern were in upper case.
3290 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3292 4. Make PCRE_ANCHORED public and accept as a compile option.
3294 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3295 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3296 pass them.
3298 6. Give an error if bad option bits passed at compile or run time.
3300 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3301 pcretest to cause it to pass that flag.
3303 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3304 options, and the first character, if set.
3306 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3309 Version 0.91 10-Sep-97
3310 ----------------------
3312 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3313 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3315 2. PCRE was looping on encountering an indefinitely repeated back reference to
3316 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3317 Perl does - treats the match as successful.
3319 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5