ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 360 - (show annotations)
Wed Jul 9 20:00:28 2008 UTC (7 years, 2 months ago) by ph10
File size: 158348 byte(s)
Error occurred while calculating annotation data.
Fix CVE-2008-2371 (outer level option with alternatives caused crash).
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).
14 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
15 scripts.
17 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
18 a group with a zero qualifier. The result of the study could be incorrect,
19 or the function might crash, depending on the pattern.
21 4. Caseless matching was not working for non-ASCII characters in back
22 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
23 It now works when Unicode Property Support is available.
25 5. In pcretest, an escape such as \x{de} in the data was always generating
26 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
27 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
28 truncation.
30 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
32 7. Added two (int) casts to pcregrep when printing the difference of two
33 pointers, in case they are 64-bit values.
35 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
36 test 2 if it fails.
38 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
39 and a #define of that name to empty if it is not externally set. This is to
40 allow users of MSVC to set it if necessary.
42 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
43 the convenience functions in the pcre_get.c source file.
45 11. An option change at the start of a pattern that had top-level alternatives
46 could cause overwriting and/or a crash. This command provoked a crash in
47 some environments:
49 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
51 This potential security problem was recorded as CVE-2008-2371.
54 Version 7.7 07-May-08
55 ---------------------
57 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
58 a string to a long long, pretend we don't even have a long long." This is
59 done by checking for the strtoq, strtoll, and _strtoi64 functions.
61 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
62 pre-7.6 versions, which defined a global no_arg variable instead of putting
63 it in the RE class. (See also #8 below.)
65 3. Remove a line of dead code, identified by coverity and reported by Nuno
66 Lopes.
68 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
70 (1) The include/exclude patterns were being applied to the whole pathnames
71 of files, instead of just to the final components.
73 (2) If there was more than one level of directory, the subdirectories were
74 skipped unless they satisfied the include/exclude conditions. This is
75 inconsistent with GNU grep (and could even be seen as contrary to the
76 pcregrep specification - which I improved to make it absolutely clear).
77 The action now is always to scan all levels of directory, and just
78 apply the include/exclude patterns to regular files.
80 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
81 --exclude_dir in the tests to avoid scanning .svn directories.
83 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
84 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
85 doesn't support NULs in patterns.
87 7. Added some missing "const"s to declarations of static tables in
88 pcre_compile.c and pcre_dfa_exec.c.
90 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
91 caused by fix #2 above. (Subsequently also a second patch to fix the
92 first patch. And a third patch - this was a messy problem.)
94 9. Applied Craig's patch to remove the use of push_back().
96 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
97 matching function regexec().
99 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
100 which, however, unlike Perl's \g{...}, are subroutine calls, not back
101 references. PCRE supports relative numbers with this syntax (I don't think
102 Oniguruma does).
104 12. Previously, a group with a zero repeat such as (...){0} was completely
105 omitted from the compiled regex. However, this means that if the group
106 was called as a subroutine from elsewhere in the pattern, things went wrong
107 (an internal error was given). Such groups are now left in the compiled
108 pattern, with a new opcode that causes them to be skipped at execution
109 time.
111 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
112 to the way PCRE behaves:
114 (a) A lone ] character is dis-allowed (Perl treats it as data).
116 (b) A back reference to an unmatched subpattern matches an empty string
117 (Perl fails the current match path).
119 (c) A data ] in a character class must be notated as \] because if the
120 first data character in a class is ], it defines an empty class. (In
121 Perl it is not possible to have an empty class.) The empty class []
122 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
123 The negative empty class [^] matches any one character, independently
124 of the DOTALL setting.
126 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
127 non-existent subpattern following a character class starting with ']' and
128 containing () gave an internal compiling error instead of "reference to
129 non-existent subpattern". Fortunately, when the pattern did exist, the
130 compiled code was correct. (When scanning forwards to check for the
131 existencd of the subpattern, it was treating the data ']' as terminating
132 the class, so got the count wrong. When actually compiling, the reference
133 was subsequently set up correctly.)
135 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
136 it was being rejected as not supported by pcre_dfa_exec(), even though
137 other assertions are supported. I have made pcre_dfa_exec() support
138 (*FAIL).
140 16. The implementation of 13c above involved the invention of a new opcode,
141 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
142 cannot be changed at match time, I realized I could make a small
143 improvement to matching performance by compiling OP_ALLANY instead of
144 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
145 on the OP_ANY path.
147 17. Compiling pcretest on Windows with readline support failed without the
148 following two fixes: (1) Make the unistd.h include conditional on
149 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
151 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
152 ncurses library to be included for pcretest when ReadLine support is
153 requested, but also to allow for it to be overridden. This patch came from
154 Daniel Bergström.
156 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
157 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
158 any errors with the current Unicode tables. Thanks to Peter Kankowski for
159 spotting this.
162 Version 7.6 28-Jan-08
163 ---------------------
165 1. A character class containing a very large number of characters with
166 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
167 overflow.
169 2. Patch to cut out the "long long" test in pcrecpp_unittest when
170 HAVE_LONG_LONG is not defined.
172 3. Applied Christian Ehrlicher's patch to update the CMake build files to
173 bring them up to date and include new features. This patch includes:
175 - Fixed PH's badly added libz and libbz2 support.
176 - Fixed a problem with static linking.
177 - Added pcredemo. [But later removed - see 7 below.]
178 - Fixed dftables problem and added an option.
179 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
181 - Added readline support for pcretest.
182 - Added an listing of the option settings after cmake has run.
184 4. A user submitted a patch to Makefile that makes it easy to create
185 "pcre.dll" under mingw when using Configure/Make. I added stuff to
186 Makefile.am that cause it to include this special target, without
187 affecting anything else. Note that the same mingw target plus all
188 the other distribution libraries and programs are now supported
189 when configuring with CMake (see 6 below) instead of with
190 Configure/Make.
192 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
193 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
194 exported in the Windows port". It has not yet been confirmed that the patch
195 solves the problem, but it does no harm.
197 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
198 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
199 with CMake, and also correct the comment about stack recursion.
201 7. Remove the automatic building of pcredemo from the ./configure system and
202 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
203 of a program that users should build themselves after PCRE is installed, so
204 building it automatically is not really right. What is more, it gave
205 trouble in some build environments.
207 8. Further tidies to CMakeLists.txt from Sheri and Christian.
210 Version 7.5 10-Jan-08
211 ---------------------
213 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
214 values in parens when parsing an RE using the C++ wrapper."
216 2. Negative specials like \S did not work in character classes in UTF-8 mode.
217 Characters greater than 255 were excluded from the class instead of being
218 included.
220 3. The same bug as (2) above applied to negated POSIX classes such as
221 [:^space:].
223 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
224 defined or documented. It seems to have been a typo for PCRE_STATIC, so
225 I have changed it.
227 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
228 first named subpattern) and a construct such as (?&a) would reference the
229 first named subpattern whose name started with "a" (in other words, the
230 length check was missing). Both these problems are fixed. "Subpattern name
231 expected" is now given for (?&) (a zero-length name), and this patch also
232 makes it give the same error for \k'' (previously it complained that that
233 was a reference to a non-existent subpattern).
235 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
236 this is right because (?- can be followed by option settings as well as by
237 digits. I have, however, made the messages clearer.
239 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
240 than the number used in the conditional) now cause a compile-time error.
241 This is actually not compatible with Perl, which accepts such patterns, but
242 treats the conditional as always being FALSE (as PCRE used to), but it
243 seems to me that giving a diagnostic is better.
245 8. Change "alphameric" to the more common word "alphanumeric" in comments
246 and messages.
248 9. Fix two occurrences of "backslash" in comments that should have been
249 "backspace".
251 10. Remove two redundant lines of code that can never be obeyed (their function
252 was moved elsewhere).
254 11. The program that makes PCRE's Unicode character property table had a bug
255 which caused it to generate incorrect table entries for sequences of
256 characters that have the same character type, but are in different scripts.
257 It amalgamated them into a single range, with the script of the first of
258 them. In other words, some characters were in the wrong script. There were
259 thirteen such cases, affecting characters in the following ranges:
261 U+002b0 - U+002c1
262 U+0060c - U+0060d
263 U+0061e - U+00612
264 U+0064b - U+0065e
265 U+0074d - U+0076d
266 U+01800 - U+01805
267 U+01d00 - U+01d77
268 U+01d9b - U+01dbf
269 U+0200b - U+0200f
270 U+030fc - U+030fe
271 U+03260 - U+0327f
272 U+0fb46 - U+0fbb1
273 U+10450 - U+1049d
275 12. The -o option (show only the matching part of a line) for pcregrep was not
276 compatible with GNU grep in that, if there was more than one match in a
277 line, it showed only the first of them. It now behaves in the same way as
278 GNU grep.
280 13. If the -o and -v options were combined for pcregrep, it printed a blank
281 line for every non-matching line. GNU grep prints nothing, and pcregrep now
282 does the same. The return code can be used to tell if there were any
283 non-matching lines.
285 14. Added --file-offsets and --line-offsets to pcregrep.
287 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
288 infinitely looping recursion. The bug was that positive lookaheads were not
289 being skipped when checking for a possible empty match (negative lookaheads
290 and both kinds of lookbehind were skipped).
292 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
293 inclusion of <windows.h> to before rather than after the definition of
294 INVALID_FILE_ATTRIBUTES (patch from David Byron).
296 17. Specifying a possessive quantifier with a specific limit for a Unicode
297 character property caused pcre_compile() to compile bad code, which led at
298 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
299 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
300 caused the error; without that there was no problem.
302 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
304 19. Added --enable-pcretest-libreadline.
306 20. In pcrecpp.cc, the variable 'count' was incremented twice in
307 RE::GlobalReplace(). As a result, the number of replacements returned was
308 double what it should be. I removed one of the increments, but Craig sent a
309 later patch that removed the other one (the right fix) and added unit tests
310 that check the return values (which was not done before).
312 21. Several CMake things:
314 (1) Arranged that, when cmake is used on Unix, the libraries end up with
315 the names libpcre and libpcreposix, not just pcre and pcreposix.
317 (2) The above change means that pcretest and pcregrep are now correctly
318 linked with the newly-built libraries, not previously installed ones.
322 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
323 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
324 UTF-8 newline character). The key issue is that the pattern starts .*;
325 this means that the match must be either at the beginning, or after a
326 newline. The bug was in the code for advancing after a failed match and
327 checking that the new position followed a newline. It was not taking
328 account of UTF-8 characters correctly.
330 23. PCRE was behaving differently from Perl in the way it recognized POSIX
331 character classes. PCRE was not treating the sequence [:...:] as a
332 character class unless the ... were all letters. Perl, however, seems to
333 allow any characters between [: and :], though of course it rejects as
334 unknown any "names" that contain non-letters, because all the known class
335 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
336 for example, whereas PCRE did not - it did not recognize a POSIX character
337 class. This seemed a bit dangerous, so the code has been changed to be
338 closer to Perl. The behaviour is not identical to Perl, because PCRE will
339 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
340 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
341 Perl does, and where it didn't before.
343 24. Rewrite so as to remove the single use of %n from pcregrep because in some
344 Windows environments %n is disabled by default.
347 Version 7.4 21-Sep-07
348 ---------------------
350 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
351 means that a class such as [\s] counted as "explicit reference to CR or
352 LF". That isn't really right - the whole point of the change was to try to
353 help when there was an actual mention of one of the two characters. So now
354 the change happens only if \r or \n (or a literal CR or LF) character is
355 encountered.
357 2. The 32-bit options word was also used for 6 internal flags, but the numbers
358 of both had grown to the point where there were only 3 bits left.
359 Fortunately, there was spare space in the data structure, and so I have
360 moved the internal flags into a new 16-bit field to free up more option
361 bits.
363 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
364 but did not set the internal JCHANGED flag - either of these is enough to
365 control the way the "get" function works - but the PCRE_INFO_JCHANGED
366 facility is supposed to tell if (?J) was ever used, so now (?J) at the
367 start sets both bits.
369 4. Added options (at build time, compile time, exec time) to change \R from
370 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
372 5. doc/pcresyntax.html was missing from the distribution.
374 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
375 compatibility, even though it is no longer used.
377 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
378 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
379 windows.h file is present (where different names are used). [This was
380 reversed later after testing - see 16 below.]
382 8. Changed all #include <config.h> to #include "config.h". There were also
383 some further <pcre.h> cases that I changed to "pcre.h".
385 9. When pcregrep was used with the --colour option, it missed the line ending
386 sequence off the lines that it output.
388 10. It was pointed out to me that arrays of string pointers cause lots of
389 relocations when a shared library is dynamically loaded. A technique of
390 using a single long string with a table of offsets can drastically reduce
391 these. I have refactored PCRE in four places to do this. The result is
392 dramatic:
394 Originally: 290
395 After changing UCP table: 187
396 After changing error message table: 43
397 After changing table of "verbs" 36
398 After changing table of Posix names 22
400 Thanks to the folks working on Gregex for glib for this insight.
402 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
403 unicode-properties was also set.
405 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
407 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
408 checked only for CRLF.
410 14. Added casts to pcretest.c to avoid compiler warnings.
412 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
414 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
415 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
416 entirely. This removes changes made in 7 above.
418 17. The CMake files have been updated, and there is now more information about
419 building with CMake in the NON-UNIX-USE document.
422 Version 7.3 28-Aug-07
423 ---------------------
425 1. In the rejigging of the build system that eventually resulted in 7.1, the
426 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
427 brackets there is not right, since it causes compilers to look for an
428 installed pcre.h, not the version that is in the source that is being
429 compiled (which of course may be different). I have changed it back to:
431 #include "pcre.h"
433 I have a vague recollection that the change was concerned with compiling in
434 different directories, but in the new build system, that is taken care of
435 by the VPATH setting the Makefile.
437 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
438 when the subject happened to end in the byte 0x85 (e.g. if the last
439 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
440 characters but of course it shouldn't be taken as a newline when it is part
441 of another character. The bug was that, for an unlimited repeat of . in
442 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
443 characters when looking for a newline.
445 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
447 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
448 in debug output.
450 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
451 long printing in the pcrecpp unittest when running under MinGW.
453 6. ESC_K was left out of the EBCDIC table.
455 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
456 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
457 limit also applies to "virtual nesting" when a pattern is recursive, and in
458 this case 1000 isn't so big. I have been able to remove this limit at the
459 expense of backing off one optimization in certain circumstances. Normally,
460 when pcre_exec() would call its internal match() function recursively and
461 immediately return the result unconditionally, it uses a "tail recursion"
462 feature to save stack. However, when a subpattern that can match an empty
463 string has an unlimited repetition quantifier, it no longer makes this
464 optimization. That gives it a stack frame in which to save the data for
465 checking that an empty string has been matched. Previously this was taken
466 from the 1000-entry workspace that had been reserved. So now there is no
467 explicit limit, but more stack is used.
469 8. Applied Daniel's patches to solve problems with the import/export magic
470 syntax that is required for Windows, and which was going wrong for the
471 pcreposix and pcrecpp parts of the library. These were overlooked when this
472 problem was solved for the main library.
474 9. There were some crude static tests to avoid integer overflow when computing
475 the size of patterns that contain repeated groups with explicit upper
476 limits. As the maximum quantifier is 65535, the maximum group length was
477 set at 30,000 so that the product of these two numbers did not overflow a
478 32-bit integer. However, it turns out that people want to use groups that
479 are longer than 30,000 bytes (though not repeat them that many times).
480 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
481 made it possible to implement the integer overflow checks in a much more
482 dynamic way, which I have now done. The artificial limitation on group
483 length has been removed - we now have only the limit on the total length of
484 the compiled pattern, which depends on the LINK_SIZE setting.
486 10. Fixed a bug in the documentation for get/copy named substring when
487 duplicate names are permitted. If none of the named substrings are set, the
488 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
489 empty string.
491 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
492 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
493 because the ] is interpreted as the first data character and the
494 terminating ] is not found. PCRE has been made compatible with Perl in this
495 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
496 cause memory overwriting.
498 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
499 string has been matched (to stop an infinite loop). It was not recognizing
500 a conditional subpattern that could match an empty string if that
501 subpattern was within another subpattern. For example, it looped when
502 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
503 condition was not nested. This bug has been fixed.
505 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
506 past the start of the subject in the presence of bytes with the top bit
507 set, for example "\x8aBCD".
509 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
510 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
512 14. Optimized (?!) to (*FAIL).
514 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
515 This restricts code points to be within the range 0 to 0x10FFFF, excluding
516 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
517 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
518 does: it's just the validity check that is more restrictive.
520 16. Inserted checks for integer overflows during escape sequence (backslash)
521 processing, and also fixed erroneous offset values for syntax errors during
522 backslash processing.
524 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
525 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
527 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
528 caused an overrun.
530 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
531 something other than just ASCII characters) inside a group that had an
532 unlimited repeat caused a loop at compile time (while checking to see
533 whether the group could match an empty string).
535 20. Debugging a pattern containing \p or \P could cause a crash. For example,
536 [\P{Any}] did so. (Error in the code for printing property names.)
538 21. An orphan \E inside a character class could cause a crash.
540 22. A repeated capturing bracket such as (A)? could cause a wild memory
541 reference during compilation.
543 23. There are several functions in pcre_compile() that scan along a compiled
544 expression for various reasons (e.g. to see if it's fixed length for look
545 behind). There were bugs in these functions when a repeated \p or \P was
546 present in the pattern. These operators have additional parameters compared
547 with \d, etc, and these were not being taken into account when moving along
548 the compiled data. Specifically:
550 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
551 length.
553 (b) An item such as \pL+ within a repeated group could cause crashes or
554 loops.
556 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
557 "reference to non-existent subpattern" error.
559 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
561 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
562 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
564 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
566 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
567 character were causing crashes (broken optimization).
569 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
570 \p or \P) caused a compile-time loop.
572 28. More problems have arisen in unanchored patterns when CRLF is a valid line
573 break. For example, the unstudied pattern [\r\n]A does not match the string
574 "\r\nA" because change 7.0/46 below moves the current point on by two
575 characters after failing to match at the start. However, the pattern \nA
576 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
577 the same is true. There doesn't seem any very clean way out of this, but
578 what I have chosen to do makes the common cases work: PCRE now takes note
579 of whether there can be an explicit match for \r or \n anywhere in the
580 pattern, and if so, 7.0/46 no longer applies. As part of this change,
581 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
582 pattern has explicit CR or LF references.
584 29. Added (*CR) etc for changing newline setting at start of pattern.
587 Version 7.2 19-Jun-07
588 ---------------------
590 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
591 which is apparently normally available under Windows.
593 2. Re-jig the pcregrep tests with different newline settings in an attempt
594 to make them independent of the local environment's newline setting.
596 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
598 4. Some of the "internals" tests were previously cut out when the link size
599 was not 2, because the output contained actual offsets. The recent new
600 "Z" feature of pcretest means that these can be cut out, making the tests
601 usable with all link sizes.
603 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
604 stack recursion. This gives a massive performance boost under BSD, but just
605 a small improvement under Linux. However, it saves one field in the frame
606 in all cases.
608 6. Added more features from the forthcoming Perl 5.10:
610 (a) (?-n) (where n is a string of digits) is a relative subroutine or
611 recursion call. It refers to the nth most recently opened parentheses.
613 (b) (?+n) is also a relative subroutine call; it refers to the nth next
614 to be opened parentheses.
616 (c) Conditions that refer to capturing parentheses can be specified
617 relatively, for example, (?(-2)... or (?(+3)...
619 (d) \K resets the start of the current match so that everything before
620 is not part of it.
622 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
624 (f) \g{name} is another synonym - part of Perl 5.10's unification of
625 reference syntax.
627 (g) (?| introduces a group in which the numbering of parentheses in each
628 alternative starts with the same number.
630 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
632 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
635 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
636 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
637 for detecting groups that can match an empty string.
639 9. A pattern with a very large number of alternatives (more than several
640 hundred) was running out of internal workspace during the pre-compile
641 phase, where pcre_compile() figures out how much memory will be needed. A
642 bit of new cunning has reduced the workspace needed for groups with
643 alternatives. The 1000-alternative test pattern now uses 12 bytes of
644 workspace instead of running out of the 4096 that are available.
646 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
648 11. Applied patch from Google to remove an optimization that didn't quite work.
649 The report of the bug said:
651 pcrecpp::RE("a*").FullMatch("aaa") matches, while
652 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
653 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
655 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
656 it matched the wrong number of bytes.
659 Version 7.1 24-Apr-07
660 ---------------------
662 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
663 that is more "standard", making use of automake and other Autotools. There
664 is some re-arrangement of the files and adjustment of comments consequent
665 on this.
667 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
668 for recursive directory scanning broke on some systems because the files
669 are not scanned in any specific order and on different systems the order
670 was different. A call to "sort" has been inserted into RunGrepTest for the
671 approprate test as a short-term fix. In the longer term there may be an
672 alternative.
674 3. I had an email from Eric Raymond about problems translating some of PCRE's
675 man pages to HTML (despite the fact that I distribute HTML pages, some
676 people do their own conversions for various reasons). The problems
677 concerned the use of low-level troff macros .br and .in. I have therefore
678 removed all such uses from the man pages (some were redundant, some could
679 be replaced by .nf/.fi pairs). The 132html script that I use to generate
680 HTML has been updated to handle .nf/.fi and to complain if it encounters
681 .br or .in.
683 4. Updated comments in configure.ac that get placed in config.h.in and also
684 arranged for config.h to be included in the distribution, with the name
685 config.h.generic, for the benefit of those who have to compile without
686 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
688 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
689 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
690 makevp.bat for the current PCRE, using the additional files
691 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
693 6. A Windows user reported a minor discrepancy with test 2, which turned out
694 to be caused by a trailing space on an input line that had got lost in his
695 copy. The trailing space was an accident, so I've just removed it.
697 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
698 that is needed.
700 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
701 as "const" (a) because they are and (b) because it helps the PHP
702 maintainers who have recently made a script to detect big data structures
703 in the php code that should be moved to the .rodata section. I remembered
704 to update Builducptable as well, so it won't revert if ucptable.h is ever
705 re-created.
707 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
708 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
709 order to be able to cut out the UTF-8 tables in the latter when UTF-8
710 support is not required. This saves 1.5-2K of code, which is important in
711 some applications.
713 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
714 so as not to refer to the tables, even though these functions will never be
715 called when UTF-8 support is disabled. Otherwise there are problems with a
716 shared library.
718 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
720 (a) It was defining its arguments as char * instead of void *.
722 (b) It was assuming that all moves were upwards in memory; this was true
723 a long time ago when I wrote it, but is no longer the case.
725 The emulated memove() is provided for those environments that have neither
726 memmove() nor bcopy(). I didn't think anyone used it these days, but that
727 is clearly not the case, as these two bugs were recently reported.
729 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
730 and Detrail to create the HTML documentation, the .txt form of the man
731 pages, and it removes trailing spaces from listed files. It also creates
732 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
733 case, it wraps all the #defines with #ifndefs. This script should be run
734 before "make dist".
736 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
737 with Unicode property support.
739 (a) For a maximizing quantifier, if the two different cases of the
740 character were of different lengths in their UTF-8 codings (there are
741 some cases like this - I found 11), and the matching function had to
742 back up over a mixture of the two cases, it incorrectly assumed they
743 were both the same length.
745 (b) When PCRE was configured to use the heap rather than the stack for
746 recursion during matching, it was not correctly preserving the data for
747 the other case of a UTF-8 character when checking ahead for a match
748 while processing a minimizing repeat. If the check also involved
749 matching a wide character, but failed, corruption could cause an
750 erroneous result when trying to check for a repeat of the original
751 character.
753 13. Some tidying changes to the testing mechanism:
755 (a) The RunTest script now detects the internal link size and whether there
756 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
757 values substituted by "configure". (The RunGrepTest script already did
758 this for UTF-8.) The configure.ac script no longer substitutes the
759 relevant variables.
761 (b) The debugging options /B and /D in pcretest show the compiled bytecode
762 with length and offset values. This means that the output is different
763 for different internal link sizes. Test 2 is skipped for link sizes
764 other than 2 because of this, bypassing the problem. Unfortunately,
765 there was also a test in test 3 (the locale tests) that used /B and
766 failed for link sizes other than 2. Rather than cut the whole test out,
767 I have added a new /Z option to pcretest that replaces the length and
768 offset values with spaces. This is now used to make test 3 independent
769 of link size. (Test 2 will be tidied up later.)
771 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
772 segmentation fault instead of returning the appropriate error message.
774 15. In multiline mode when the newline sequence was set to "any", the pattern
775 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
776 This doesn't seem right; it now treats the CRLF combination as the line
777 ending, and so does not match in that case. It's only a pattern such as ^$
778 that would hit this one: something like ^ABC$ would have failed after \r
779 and then tried again after \r\n.
781 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
782 in an attempt to make files that differ only in their line terminators
783 compare equal. This works on Linux.
785 17. Under certain error circumstances pcregrep might try to free random memory
786 as it exited. This is now fixed, thanks to valgrind.
788 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
789 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
790 was because its rules for how to advance for /g after matching an empty
791 string at the end of a line did not allow for this case. They now check for
792 it specially.
794 20. pcretest is supposed to handle patterns and data of any length, by
795 extending its buffers when necessary. It was getting this wrong when the
796 buffer for a data line had to be extended.
798 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
799 CRLF as a newline sequence.
801 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
802 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
803 I have nevertheless tidied it up.
805 23. Added some casts to kill warnings from HP-UX ia64 compiler.
807 24. Added a man page for pcre-config.
810 Version 7.0 19-Dec-06
811 ---------------------
813 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
814 moving to gcc 4.1.1.
816 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
817 sys/time.h, which is documented as needed for this function. It doesn't
818 seem to matter on Linux, but it showed up on some releases of OS X.
820 3. It seems that there are systems where bytes whose values are greater than
821 127 match isprint() in the "C" locale. The "C" locale should be the
822 default when a C program starts up. In most systems, only ASCII printing
823 characters match isprint(). This difference caused the output from pcretest
824 to vary, making some of the tests fail. I have changed pcretest so that:
826 (a) When it is outputting text in the compiled version of a pattern, bytes
827 other than 32-126 are always shown as hex escapes.
829 (b) When it is outputting text that is a matched part of a subject string,
830 it does the same, unless a different locale has been set for the match
831 (using the /L modifier). In this case, it uses isprint() to decide.
833 4. Fixed a major bug that caused incorrect computation of the amount of memory
834 required for a compiled pattern when options that changed within the
835 pattern affected the logic of the preliminary scan that determines the
836 length. The relevant options are -x, and -i in UTF-8 mode. The result was
837 that the computed length was too small. The symptoms of this bug were
838 either the PCRE error "internal error: code overflow" from pcre_compile(),
839 or a glibc crash with a message such as "pcretest: free(): invalid next
840 size (fast)". Examples of patterns that provoked this bug (shown in
841 pcretest format) are:
843 /(?-x: )/x
844 /(?x)(?-x: \s*#\s*)/
845 /((?i)[\x{c0}])/8
846 /(?i:[\x{c0}])/8
848 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
849 is now done differently.
851 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
852 wrapper classes; (b) implement a new function in the C++ scanner that is
853 more efficient than the old way of doing things because it avoids levels of
854 recursion in the regex matching; (c) add a paragraph to the documentation
855 for the FullMatch() function.
857 6. The escape sequence \n was being treated as whatever was defined as
858 "newline". Not only was this contrary to the documentation, which states
859 that \n is character 10 (hex 0A), but it also went horribly wrong when
860 "newline" was defined as CRLF. This has been fixed.
862 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
863 was being set to -1 for the "end of line" case (supposedly a value that no
864 character can have). Though this value is never used (the check for end of
865 line is "zero bytes in current character"), it caused compiler complaints.
866 I've changed it to 0xffffffff.
868 8. In pcre_version.c, the version string was being built by a sequence of
869 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
870 string (as it is for production releases) called a macro with an empty
871 argument. The C standard says the result of this is undefined. The gcc
872 compiler treats it as an empty string (which was what was wanted) but it is
873 reported that Visual C gives an error. The source has been hacked around to
874 avoid this problem.
876 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
877 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
878 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
879 of them did).
881 10. Originally, pcretest opened its input and output without "b"; then I was
882 told that "b" was needed in some environments, so it was added for release
883 5.0 to both the input and output. (It makes no difference on Unix-like
884 systems.) Later I was told that it is wrong for the input on Windows. I've
885 now abstracted the modes into two macros, to make it easier to fiddle with
886 them, and removed "b" from the input mode under Windows.
888 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
890 12. Added -help and --help to pcretest as an official way of being reminded
891 of the options.
893 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
894 and pcrecpp.cc because they annoy compilers at high warning levels.
896 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
898 15. Fixed an occurrence of == in configure.ac that should have been = (shell
899 scripts are not C programs :-) and which was not noticed because it works
900 on Linux.
902 16. pcretest is supposed to handle any length of pattern and data line (as one
903 line or as a continued sequence of lines) by extending its input buffer if
904 necessary. This feature was broken for very long pattern lines, leading to
905 a string of junk being passed to pcre_compile() if the pattern was longer
906 than about 50K.
908 17. I have done a major re-factoring of the way pcre_compile() computes the
909 amount of memory needed for a compiled pattern. Previously, there was code
910 that made a preliminary scan of the pattern in order to do this. That was
911 OK when PCRE was new, but as the facilities have expanded, it has become
912 harder and harder to keep it in step with the real compile phase, and there
913 have been a number of bugs (see for example, 4 above). I have now found a
914 cunning way of running the real compile function in a "fake" mode that
915 enables it to compute how much memory it would need, while actually only
916 ever using a few hundred bytes of working memory and without too many
917 tests of the mode. This should make future maintenance and development
918 easier. A side effect of this work is that the limit of 200 on the nesting
919 depth of parentheses has been removed (though this was never a serious
920 limitation, I suspect). However, there is a downside: pcre_compile() now
921 runs more slowly than before (30% or more, depending on the pattern). I
922 hope this isn't a big issue. There is no effect on runtime performance.
924 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
925 newline (only possible for the last line of a file) and it was a
926 pattern that set a locale (followed by /Lsomething), pcretest crashed.
928 19. Added additional timing features to pcretest. (1) The -tm option now times
929 matching only, not compiling. (2) Both -t and -tm can be followed, as a
930 separate command line item, by a number that specifies the number of
931 repeats to use when timing. The default is 50000; this gives better
932 precision, but takes uncomfortably long for very large patterns.
934 20. Extended pcre_study() to be more clever in cases where a branch of a
935 subpattern has no definite first character. For example, (a*|b*)[cd] would
936 previously give no result from pcre_study(). Now it recognizes that the
937 first character must be a, b, c, or d.
939 21. There was an incorrect error "recursive call could loop indefinitely" if
940 a subpattern (or the entire pattern) that was being tested for matching an
941 empty string contained only one non-empty item after a nested subpattern.
942 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
943 incorrectly, because the \d was being skipped in the check.
945 22. The pcretest program now has a new pattern option /B and a command line
946 option -b, which is equivalent to adding /B to every pattern. This causes
947 it to show the compiled bytecode, without the additional information that
948 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
949 is the same as /B/I).
951 23. A new optimization is now able automatically to treat some sequences such
952 as a*b as a*+b. More specifically, if something simple (such as a character
953 or a simple class like \d) has an unlimited quantifier, and is followed by
954 something that cannot possibly match the quantified thing, the quantifier
955 is automatically "possessified".
957 24. A recursive reference to a subpattern whose number was greater than 39
958 went wrong under certain circumstances in UTF-8 mode. This bug could also
959 have affected the operation of pcre_study().
961 25. Realized that a little bit of performance could be had by replacing
962 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
964 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
966 27. Possessive quantifiers such as a++ were previously implemented by turning
967 them into atomic groups such as ($>a+). Now they have their own opcodes,
968 which improves performance. This includes the automatically created ones
969 from 23 above.
971 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
972 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
973 the first matched character to be a colon. This applied both to named and
974 numbered groups.
976 29. The ucpinternal.h header file was missing its idempotency #ifdef.
978 30. I was sent a "project" file called libpcre.a.dev which I understand makes
979 building PCRE on Windows easier, so I have included it in the distribution.
981 31. There is now a check in pcretest against a ridiculously large number being
982 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
983 loop, the loop is abandoned.
985 32. Forward references to subpatterns in conditions such as (?(2)...) where
986 subpattern 2 is defined later cause pcre_compile() to search forwards in
987 the pattern for the relevant set of parentheses. This search went wrong
988 when there were unescaped parentheses in a character class, parentheses
989 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
991 33. "Subroutine" calls and backreferences were previously restricted to
992 referencing subpatterns earlier in the regex. This restriction has now
993 been removed.
995 34. Added a number of extra features that are going to be in Perl 5.10. On the
996 whole, these are just syntactic alternatives for features that PCRE had
997 previously implemented using the Python syntax or my own invention. The
998 other formats are all retained for compatibility.
1000 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1001 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1002 also .NET compatible.
1004 (b) A recursion or subroutine call to a named group can now be defined as
1005 (?&name) as well as (?P>name).
1007 (c) A backreference to a named group can now be defined as \k<name> or
1008 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1009 5.10, are also .NET compatible.
1011 (d) A conditional reference to a named group can now use the syntax
1012 (?(<name>) or (?('name') as well as (?(name).
1014 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1015 groups (named and numbered) that are never evaluated inline, but can be
1016 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1017 is always false. There may be only one alternative in such a group.
1019 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1020 as the simple (?(R). The condition is true only if the most recent
1021 recursion is that of the given number or name. It does not search out
1022 through the entire recursion stack.
1024 (g) The escape \gN or \g{N} has been added, where N is a positive or
1025 negative number, specifying an absolute or relative reference.
1027 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1028 some "unreachable code" warnings.
1030 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1031 things, this adds five new scripts.
1033 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1034 There were also incompatibilities regarding the handling of \Q..\E inside
1035 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1036 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1038 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1039 matches an empty string, and forcibly breaks the loop. There were bugs in
1040 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1041 against aaaa the result was just "a" rather than "aaaa", for example. Two
1042 separate and independent bugs (that affected different cases) have been
1043 fixed.
1045 39. Refactored the code to abolish the use of different opcodes for small
1046 capturing bracket numbers. This is a tidy that I avoided doing when I
1047 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1048 The new approach is not only tidier, it makes it possible to reduce the
1049 memory needed to fix the previous bug (38).
1051 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1052 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1053 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1054 mode.
1056 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1057 report.
1059 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1060 copy construction and assignment in the C++ wrapper.
1062 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1063 couple of bugs that could have given wrong results in the "--newline=crlf"
1064 case.
1066 44. Added a number of casts and did some reorganization of signed/unsigned int
1067 variables following suggestions from Dair Grant. Also renamed the variable
1068 "this" as "item" because it is a C++ keyword.
1070 45. Arranged for dftables to add
1072 #include "pcre_internal.h"
1074 to pcre_chartables.c because without it, gcc 4.x may remove the array
1075 definition from the final binary if PCRE is built into a static library and
1076 dead code stripping is activated.
1078 46. For an unanchored pattern, if a match attempt fails at the start of a
1079 newline sequence, and the newline setting is CRLF or ANY, and the next two
1080 characters are CRLF, advance by two characters instead of one.
1083 Version 6.7 04-Jul-06
1084 ---------------------
1086 1. In order to handle tests when input lines are enormously long, pcretest has
1087 been re-factored so that it automatically extends its buffers when
1088 necessary. The code is crude, but this _is_ just a test program. The
1089 default size has been increased from 32K to 50K.
1091 2. The code in pcre_study() was using the value of the re argument before
1092 testing it for NULL. (Of course, in any sensible call of the function, it
1093 won't be NULL.)
1095 3. The memmove() emulation function in pcre_internal.h, which is used on
1096 systems that lack both memmove() and bcopy() - that is, hardly ever -
1097 was missing a "static" storage class specifier.
1099 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1100 containing an extended class (one that cannot be represented by a bitmap
1101 because it contains high-valued characters or Unicode property items, e.g.
1102 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1103 pattern, but PCRE should not loop if you do not (it no longer does).
1104 [Detail: two cases were found: (a) a repeated subpattern containing an
1105 extended class; (b) a recursive reference to a subpattern that followed a
1106 previous extended class. It wasn't skipping over the extended class
1107 correctly when UTF-8 mode was not set.]
1109 5. A negated single-character class was not being recognized as fixed-length
1110 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1111 compile error "lookbehind assertion is not fixed length".
1113 6. The RunPerlTest auxiliary script was showing an unexpected difference
1114 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1115 write a Perl script that can interpret lines of an input file either as
1116 byte characters or as UTF-8, which is what "perltest" was being required to
1117 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1118 can't do is switch easily at run time between having the "use utf8;" pragma
1119 or not. In the end, I fudged it by using the RunPerlTest script to insert
1120 "use utf8;" explicitly for the UTF-8 tests.
1122 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1123 the end of the subject string, contrary to the documentation and to what
1124 Perl does. This was true of both matching functions. Now it matches only at
1125 the start of the subject and immediately after *internal* newlines.
1127 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1128 a pointer to an int instead of a pointer to an unsigned long int. This
1129 caused problems on 64-bit systems.
1131 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1132 instance of the 'standard' template library not being so standard".
1134 10. There was no check on the number of named subpatterns nor the maximum
1135 length of a subpattern name. The product of these values is used to compute
1136 the size of the memory block for a compiled pattern. By supplying a very
1137 long subpattern name and a large number of named subpatterns, the size
1138 computation could be caused to overflow. This is now prevented by limiting
1139 the length of names to 32 characters, and the number of named subpatterns
1140 to 10,000.
1142 11. Subpatterns that are repeated with specific counts have to be replicated in
1143 the compiled pattern. The size of memory for this was computed from the
1144 length of the subpattern and the repeat count. The latter is limited to
1145 65535, but there was no limit on the former, meaning that integer overflow
1146 could in principle occur. The compiled length of a repeated subpattern is
1147 now limited to 30,000 bytes in order to prevent this.
1149 12. Added the optional facility to have named substrings with the same name.
1151 13. Added the ability to use a named substring as a condition, using the
1152 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1153 are numbers (not recommended). Forward references are permitted.
1155 14. Added forward references in named backreferences (if you see what I mean).
1157 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1158 pattern could run off the end of the subject. For example, the pattern
1159 "(?s)(.{1,5})"8 did this with the subject "ab".
1161 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1162 PCRE_CASELESS was set when matching characters that were quantified with ?
1163 or *.
1165 17. A character class other than a single negated character that had a minimum
1166 but no maximum quantifier - for example [ab]{6,} - was not handled
1167 correctly by pce_dfa_exec(). It would match only one character.
1169 18. A valid (though odd) pattern that looked like a POSIX character
1170 class but used an invalid character after [ (for example [[,abc,]]) caused
1171 pcre_compile() to give the error "Failed: internal error: code overflow" or
1172 in some cases to crash with a glibc free() error. This could even happen if
1173 the pattern terminated after [[ but there just happened to be a sequence of
1174 letters, a binary zero, and a closing ] in the memory that followed.
1176 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1177 over the years. Originally (before any Unicode support), just the bottom 8
1178 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1179 output from "man perlunicode" includes this:
1181 The regular expression compiler produces polymorphic opcodes. That
1182 is, the pattern adapts to the data and automatically switches to
1183 the Unicode character scheme when presented with Unicode data--or
1184 instead uses a traditional byte scheme when presented with byte
1185 data.
1187 Sadly, a wide octal escape does not cause a switch, and in a string with
1188 no other multibyte characters, these octal escapes are treated as before.
1189 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1190 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1191 Unicode string.
1193 I have not perpetrated such confusion in PCRE. Up till now, it took just
1194 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1195 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1196 translate to the appropriate multibyte character.
1198 29. Applied some refactoring to reduce the number of warnings from Microsoft
1199 and Borland compilers. This has included removing the fudge introduced
1200 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1201 a warning about an unused variable.
1203 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1204 characters since release 4.0, because Perl (from release 5.004) does not.
1205 [Or at least, is documented not to: some releases seem to be in conflict
1206 with the documentation.] However, when a pattern was studied with
1207 pcre_study() and all its branches started with \s, PCRE still included VT
1208 as a possible starting character. Of course, this did no harm; it just
1209 caused an unnecessary match attempt.
1211 22. Removed a now-redundant internal flag bit that recorded the fact that case
1212 dependency changed within the pattern. This was once needed for "required
1213 byte" processing, but is no longer used. This recovers a now-scarce options
1214 bit. Also moved the least significant internal flag bit to the most-
1215 significant bit of the word, which was not previously used (hangover from
1216 the days when it was an int rather than a uint) to free up another bit for
1217 the future.
1219 23. Added support for CRLF line endings as well as CR and LF. As well as the
1220 default being selectable at build time, it can now be changed at runtime
1221 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1222 specify that it is scanning data with non-default line endings.
1224 24. Changed the definition of CXXLINK to make it agree with the definition of
1225 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1227 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1228 recursions. This makes a big different to stack usage for some patterns.
1230 26. If a subpattern containing a named recursion or subroutine reference such
1231 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1232 the space required for the compiled pattern went wrong and gave too small a
1233 value. Depending on the environment, this could lead to "Failed: internal
1234 error: code overflow at offset 49" or "glibc detected double free or
1235 corruption" errors.
1237 27. Applied patches from Google (a) to support the new newline modes and (b) to
1238 advance over multibyte UTF-8 characters in GlobalReplace.
1240 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1241 difference for some implementation of PCRE in some Windows version.
1243 29. Added some extra testing facilities to pcretest:
1245 \q<number> in a data line sets the "match limit" value
1246 \Q<number> in a data line sets the "match recursion limt" value
1247 -S <number> sets the stack size, where <number> is in megabytes
1249 The -S option isn't available for Windows.
1252 Version 6.6 06-Feb-06
1253 ---------------------
1255 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1256 in pcreposix.h. I have copied the definition from pcre.h.
1258 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1259 because pcre.h is no longer a built file.
1261 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1262 not normally included in the compiled code.
1265 Version 6.5 01-Feb-06
1266 ---------------------
1268 1. When using the partial match feature with pcre_dfa_exec(), it was not
1269 anchoring the second and subsequent partial matches at the new starting
1270 point. This could lead to incorrect results. For example, with the pattern
1271 /1234/, partially matching against "123" and then "a4" gave a match.
1273 2. Changes to pcregrep:
1275 (a) All non-match returns from pcre_exec() were being treated as failures
1276 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1277 error message is output. Some extra information is given for the
1279 probably the only errors that are likely to be caused by users (by
1280 specifying a regex that has nested indefinite repeats, for instance).
1281 If there are more than 20 of these errors, pcregrep is abandoned.
1283 (b) A binary zero was treated as data while matching, but terminated the
1284 output line if it was written out. This has been fixed: binary zeroes
1285 are now no different to any other data bytes.
1287 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1288 used to set a locale for matching. The --locale=xxxx long option has
1289 been added (no short equivalent) to specify a locale explicitly on the
1290 pcregrep command, overriding the environment variables.
1292 (d) When -B was used with -n, some line numbers in the output were one less
1293 than they should have been.
1295 (e) Added the -o (--only-matching) option.
1297 (f) If -A or -C was used with -c (count only), some lines of context were
1298 accidentally printed for the final match.
1300 (g) Added the -H (--with-filename) option.
1302 (h) The combination of options -rh failed to suppress file names for files
1303 that were found from directory arguments.
1305 (i) Added the -D (--devices) and -d (--directories) options.
1307 (j) Added the -F (--fixed-strings) option.
1309 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1311 (l) Added the --colo(u)r option.
1313 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1314 is not present by default.
1316 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1317 items such as (?R) or (?1), when the recursion could match a number of
1318 alternatives. If it matched one of the alternatives, but subsequently,
1319 outside the recursion, there was a failure, the code tried to back up into
1320 the recursion. However, because of the way PCRE is implemented, this is not
1321 possible, and the result was an incorrect result from the match.
1323 In order to prevent this happening, the specification of recursion has
1324 been changed so that all such subpatterns are automatically treated as
1325 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1327 4. I had overlooked the fact that, in some locales, there are characters for
1328 which isalpha() is true but neither isupper() nor islower() are true. In
1329 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1330 and ordfeminine) are like this. This affected the treatment of \w and \W
1331 when they appeared in character classes, but not when they appeared outside
1332 a character class. The bit map for "word" characters is now created
1333 separately from the results of isalnum() instead of just taking it from the
1334 upper, lower, and digit maps. (Plus the underscore character, of course.)
1336 5. The above bug also affected the handling of POSIX character classes such as
1337 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1338 permanent tables. Instead, the bit maps for such a class were previously
1339 created as the appropriate unions of the upper, lower, and digit bitmaps.
1340 Now they are created by subtraction from the [[:word:]] class, which has
1341 its own bitmap.
1343 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1344 It is created by subtracting the vertical space characters (\x09, \x0a,
1345 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1346 subtraction was done in the overall bitmap for a character class, meaning
1347 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1348 be recognized. This bug has been fixed.
1350 7. Patches from the folks at Google:
1352 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1353 real life, but is still worth protecting against".
1355 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1356 regular expressions".
1358 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1359 have it.
1361 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1362 "configure" and the latter not, in order to fix a problem somebody had
1363 with compiling the Arg class on HP-UX.
1365 (e) Improve the error-handling of the C++ wrapper a little bit.
1367 (f) New tests for checking recursion limiting.
1369 8. The pcre_memmove() function, which is used only if the environment does not
1370 have a standard memmove() function (and is therefore rarely compiled),
1371 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1372 returning a result (though PCRE never actually uses the result).
1374 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1375 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1376 returned instead of calling malloc() with an overflowing number that would
1377 most likely cause subsequent chaos.
1379 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1381 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1382 with this option is matched, the nmatch and pmatch options of regexec() are
1383 ignored.
1385 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1386 provided in case anyone wants to the the POSIX interface with UTF-8
1387 strings.
1389 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1390 C++ linking (needed for some HP-UX environments).
1392 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1393 (unused parameter) and in the pcre_printint() function (omitted "default"
1394 switch label when the default is to do nothing).
1396 15. Added some code to make it possible, when PCRE is compiled as a C++
1397 library, to replace subject pointers for pcre_exec() with a smart pointer
1398 class, thus making it possible to process discontinuous strings.
1400 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1401 much the same function. They were added by different people who were trying
1402 to make PCRE easy to compile on non-Unix systems. It has been suggested
1403 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1404 for compiling on Windows systems. I have therefore replaced it with
1405 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1406 defaults to "extern" for C or "extern C" for C++, which works fine on
1407 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1408 SCOPE with something explicit in config.h. In addition:
1410 (a) pcreposix.h still had just "extern" instead of either of these macros;
1411 I have replaced it with PCRE_DATA_SCOPE.
1413 (b) Functions such as _pcre_xclass(), which are internal to the library,
1414 but external in the C sense, all had PCRE_EXPORT in their definitions.
1415 This is apparently wrong for the Windows case, so I have removed it.
1416 (It makes no difference on Unix-like systems.)
1418 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1419 of recursive calls to match(). This is different to MATCH_LIMIT because
1420 that limits the total number of calls to match(), not all of which increase
1421 the depth of recursion. Limiting the recursion depth limits the amount of
1422 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1423 when PCRE is compiled, and changed at run time. A patch from Google adds
1424 this functionality to the C++ interface.
1426 18. Changes to the handling of Unicode character properties:
1428 (a) Updated the table to Unicode 4.1.0.
1430 (b) Recognize characters that are not in the table as "Cn" (undefined).
1432 (c) I revised the way the table is implemented to a much improved format
1433 which includes recognition of ranges. It now supports the ranges that
1434 are defined in UnicodeData.txt, and it also amalgamates other
1435 characters into ranges. This has reduced the number of entries in the
1436 table from around 16,000 to around 3,000, thus reducing its size
1437 considerably. I realized I did not need to use a tree structure after
1438 all - a binary chop search is just as efficient. Having reduced the
1439 number of entries, I extended their size from 6 bytes to 8 bytes to
1440 allow for more data.
1442 (d) Added support for Unicode script names via properties such as \p{Han}.
1444 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1445 matching that character.
1447 20. When matching a repeated Unicode property with a minimum greater than zero,
1448 (for example \pL{2,}), PCRE could look past the end of the subject if it
1449 reached it while seeking the minimum number of characters. This could
1450 happen only if some of the characters were more than one byte long, because
1451 there is a check for at least the minimum number of bytes.
1453 21. Refactored the implementation of \p and \P so as to be more general, to
1454 allow for more different types of property in future. This has changed the
1455 compiled form incompatibly. Anybody with saved compiled patterns that use
1456 \p or \P will have to recompile them.
1458 22. Added "Any" and "L&" to the supported property types.
1460 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1461 but give a compile time error if the value is greater than 0xff.
1463 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1464 accidentally not being installed or uninstalled.
1466 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1467 made were to insert the current release number. This seemed silly, because
1468 it made things harder for people building PCRE on systems that don't run
1469 "configure". I have turned pcre.h into a distributed file, no longer built
1470 by "configure", with the version identification directly included. There is
1471 no longer a pcre.h.in file.
1473 However, this change necessitated a change to the pcre-config script as
1474 well. It is built from pcre-config.in, and one of the substitutions was the
1475 release number. I have updated configure.ac so that ./configure now finds
1476 the release number by grepping pcre.h.
1478 26. Added the ability to run the tests under valgrind.
1481 Version 6.4 05-Sep-05
1482 ---------------------
1484 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1485 "--" to be printed when multiple files were scanned, even when none of the
1486 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1487 consider it to be a bug, and have restored the previous behaviour.
1489 2. A couple of code tidies to get rid of compiler warnings.
1491 3. The pcretest program used to cheat by referring to symbols in the library
1492 whose names begin with _pcre_. These are internal symbols that are not
1493 really supposed to be visible externally, and in some environments it is
1494 possible to suppress them. The cheating is now confined to including
1495 certain files from the library's source, which is a bit cleaner.
1497 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1498 file's purpose clearer.
1500 5. Reorganized pcre_ucp_findchar().
1503 Version 6.3 15-Aug-05
1504 ---------------------
1506 1. The file libpcre.pc.in did not have general read permission in the tarball.
1508 2. There were some problems when building without C++ support:
1510 (a) If C++ support was not built, "make install" and "make test" still
1511 tried to test it.
1513 (b) There were problems when the value of CXX was explicitly set. Some
1514 changes have been made to try to fix these, and ...
1516 (c) --disable-cpp can now be used to explicitly disable C++ support.
1518 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1519 backslash in a target when C++ was disabled. This confuses some
1520 versions of "make", apparently. Using an intermediate variable solves
1521 this. (Same for CPP_LOBJ.)
1523 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1524 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1525 necessary on certain architectures.
1527 4. Added a setting of -export-symbols-regex to the link command to remove
1528 those symbols that are exported in the C sense, but actually are local
1529 within the library, and not documented. Their names all begin with
1530 "_pcre_". This is not a perfect job, because (a) we have to except some
1531 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1532 available (and never for static libraries). I have made a note to try to
1533 find a way round (a) in the future.
1536 Version 6.2 01-Aug-05
1537 ---------------------
1539 1. There was no test for integer overflow of quantifier values. A construction
1540 such as {1111111111111111} would give undefined results. What is worse, if
1541 a minimum quantifier for a parenthesized subpattern overflowed and became
1542 negative, the calculation of the memory size went wrong. This could have
1543 led to memory overwriting.
1545 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1547 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1548 operating environments where this matters.
1550 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1551 PCRE options from within the C++ wrapper.
1553 5. Named capturing subpatterns were not being correctly counted when a pattern
1554 was compiled. This caused two problems: (a) If there were more than 100
1555 such subpatterns, the calculation of the memory needed for the whole
1556 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1557 back references of the form \12, where the number was greater than 9, were
1558 not recognized as back references, even though there were sufficient
1559 previous subpatterns.
1561 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1562 versions of gcc, e.g. 2.95.4.
1565 Version 6.1 21-Jun-05
1566 ---------------------
1568 1. There was one reference to the variable "posix" in pcretest.c that was not
1569 surrounded by "#if !defined NOPOSIX".
1571 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1572 the cross-check on the old pcre_info() function, for the benefit of the
1573 cut-down version of PCRE that is currently imported into Exim.
1575 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1576 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1577 patterns that start (?i) but I don't think that matters. The use of (?i) is
1578 just an example; this all applies to the other options as well.
1580 4. Since libtool seems to echo the compile commands it is issuing, the output
1581 from "make" can be reduced a bit by putting "@" in front of each libtool
1582 compile command.
1584 5. Patch from the folks at Google for configure.in to be a bit more thorough
1585 in checking for a suitable C++ installation before trying to compile the
1586 C++ stuff. This should fix a reported problem when a compiler was present,
1587 but no suitable headers.
1589 6. The man pages all had just "PCRE" as their title. I have changed them to
1590 be the relevant file name. I have also arranged that these names are
1591 retained in the file doc/pcre.txt, which is a concatenation in text format
1592 of all the man pages except the little individual ones for each function.
1594 7. The NON-UNIX-USE file had not been updated for the different set of source
1595 files that come with release 6. I also added a few comments about the C++
1596 wrapper.
1599 Version 6.0 07-Jun-05
1600 ---------------------
1602 1. Some minor internal re-organization to help with my DFA experiments.
1604 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1605 didn't matter for the library itself when fully configured, but did matter
1606 when compiling without UCP support, or within Exim, where the ucp files are
1607 not imported.
1609 3. Refactoring of the library code to split up the various functions into
1610 different source modules. The addition of the new DFA matching code (see
1611 below) to a single monolithic source would have made it really too
1612 unwieldy, quite apart from causing all the code to be include in a
1613 statically linked application, when only some functions are used. This is
1614 relevant even without the DFA addition now that patterns can be compiled in
1615 one application and matched in another.
1617 The downside of splitting up is that there have to be some external
1618 functions and data tables that are used internally in different modules of
1619 the library but which are not part of the API. These have all had their
1620 names changed to start with "_pcre_" so that they are unlikely to clash
1621 with other external names.
1623 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1624 a different (DFA) algorithm. Although it is slower than the original
1625 function, it does have some advantages for certain types of matching
1626 problem.
1628 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1629 including restarting after a partial match.
1631 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1632 defined when compiling for Windows was sent to me. I have put it into the
1633 code, though I have no means of testing or verifying it.
1635 7. Added the pcre_refcount() auxiliary function.
1637 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1638 match before or at the first newline in the subject string. In pcretest,
1639 the /f option on a pattern can be used to set this.
1641 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1642 would behave wrongly. This has been present in PCRE since release 4.0.
1644 10. A number of changes to the pcregrep command:
1646 (a) Refactored how -x works; insert ^(...)$ instead of setting
1647 PCRE_ANCHORED and checking the length, in preparation for adding
1648 something similar for -w.
1650 (b) Added the -w (match as a word) option.
1652 (c) Refactored the way lines are read and buffered so as to have more
1653 than one at a time available.
1655 (d) Implemented a pcregrep test script.
1657 (e) Added the -M (multiline match) option. This allows patterns to match
1658 over several lines of the subject. The buffering ensures that at least
1659 8K, or the rest of the document (whichever is the shorter) is available
1660 for matching (and similarly the previous 8K for lookbehind assertions).
1662 (f) Changed the --help output so that it now says
1664 -w, --word-regex(p)
1666 instead of two lines, one with "regex" and the other with "regexp"
1667 because that confused at least one person since the short forms are the
1668 same. (This required a bit of code, as the output is generated
1669 automatically from a table. It wasn't just a text change.)
1671 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1672 option but starts with a hyphen. Could be a pattern or a path name
1673 starting with a hyphen, for instance.
1675 (h) "-" can be given as a file name to represent stdin.
1677 (i) When file names are being printed, "(standard input)" is used for
1678 the standard input, for compatibility with GNU grep. Previously
1679 "<stdin>" was used.
1681 (j) The option --label=xxx can be used to supply a name to be used for
1682 stdin when file names are being printed. There is no short form.
1684 (k) Re-factored the options decoding logic because we are going to add
1685 two more options that take data. Such options can now be given in four
1686 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1688 (l) Added the -A, -B, and -C options for requesting that lines of context
1689 around matches be printed.
1691 (m) Added the -L option to print the names of files that do not contain
1692 any matching lines, that is, the complement of -l.
1694 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1695 continue to scan other files.
1697 (o) The -s option was incorrectly implemented. For compatibility with other
1698 greps, it now suppresses the error message for a non-existent or non-
1699 accessible file (but not the return code). There is a new option called
1700 -q that suppresses the output of matching lines, which was what -s was
1701 previously doing.
1703 (p) Added --include and --exclude options to specify files for inclusion
1704 and exclusion when recursing.
1706 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1707 Hopefully, it now does.
1709 12. Missing cast in pcre_study().
1711 13. Added an "uninstall" target to the makefile.
1713 14. Replaced "extern" in the function prototypes in Makefile.in with
1714 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1715 world, but is set differently for Windows.
1717 15. Added a second compiling function called pcre_compile2(). The only
1718 difference is that it has an extra argument, which is a pointer to an
1719 integer error code. When there is a compile-time failure, this is set
1720 non-zero, in addition to the error test pointer being set to point to an
1721 error message. The new argument may be NULL if no error number is required
1722 (but then you may as well call pcre_compile(), which is now just a
1723 wrapper). This facility is provided because some applications need a
1724 numeric error indication, but it has also enabled me to tidy up the way
1725 compile-time errors are handled in the POSIX wrapper.
1727 16. Added VPATH=.libs to the makefile; this should help when building with one
1728 prefix path and installing with another. (Or so I'm told by someone who
1729 knows more about this stuff than I do.)
1731 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1732 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1733 match everything, including newlines. This is not POSIX-compatible, but
1734 somebody wanted the feature. From pcretest it can be activated by using
1735 both the P and the s flags.
1737 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1739 19. libpcre.pc was being incorrectly installed as executable.
1741 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1742 it now also looks for '\r' so that it will work unmodified on Windows.
1744 21. Added Google's contributed C++ wrapper to the distribution.
1746 22. Added some untidy missing memory free() calls in pcretest, to keep
1747 Electric Fence happy when testing.
1751 Version 5.0 13-Sep-04
1752 ---------------------
1754 1. Internal change: literal characters are no longer packed up into items
1755 containing multiple characters in a single byte-string. Each character
1756 is now matched using a separate opcode. However, there may be more than one
1757 byte in the character in UTF-8 mode.
1759 2. The pcre_callout_block structure has two new fields: pattern_position and
1760 next_item_length. These contain the offset in the pattern to the next match
1761 item, and its length, respectively.
1763 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1764 insertion of callouts before each pattern item. Added the /C option to
1765 pcretest to make use of this.
1767 4. On the advice of a Windows user, the lines
1769 #if defined(_WIN32) || defined(WIN32)
1770 _setmode( _fileno( stdout ), 0x8000 );
1771 #endif /* defined(_WIN32) || defined(WIN32) */
1773 have been added to the source of pcretest. This apparently does useful
1774 magic in relation to line terminators.
1776 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1777 for the benefit of those environments where the "b" makes a difference.
1779 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1780 to know about it. I have put a hack into configure.in that adds in code
1781 to set GCC=yes if CC=icc. This seems to end up at a point in the
1782 generated configure script that is early enough to affect the setting of
1783 compiler options, which is what is needed, but I have no means of testing
1784 whether it really works. (The user who reported this had patched the
1785 generated configure script, which of course I cannot do.)
1787 LATER: After change 22 below (new libtool files), the configure script
1788 seems to know about icc (and also ecc). Therefore, I have commented out
1789 this hack in configure.in.
1791 7. Added support for pkg-config (2 patches were sent in).
1793 8. Negated POSIX character classes that used a combination of internal tables
1794 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1795 [[:^ascii]]. Typically, they would match almost any characters. The other
1796 POSIX classes were not broken in this way.
1798 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1799 to find the match, as PCRE was deluded into thinking that the match had to
1800 start at the start point or following a newline. The same bug applied to
1801 patterns with negative forward assertions or any backward assertions
1802 preceding ".*" at the start, unless the pattern required a fixed first
1803 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1805 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1806 starting at the last subject character, bytes beyond the end of the subject
1807 string were read.
1809 11. Renamed the variable "class" as "classbits" to make life easier for C++
1810 users. (Previously there was a macro definition, but it apparently wasn't
1811 enough.)
1813 12. Added the new field "tables" to the extra data so that tables can be passed
1814 in at exec time, or the internal tables can be re-selected. This allows
1815 a compiled regex to be saved and re-used at a later time by a different
1816 program that might have everything at different addresses.
1818 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1819 -R library as well as a -L library.
1821 14. The debugging options of pcretest (-d on the command line or D on a
1822 pattern) showed incorrect output for anything following an extended class
1823 that contained multibyte characters and which was followed by a quantifier.
1825 15. Added optional support for general category Unicode character properties
1826 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1827 support. It adds about 90K to the size of the library. The meanings of the
1828 inbuilt class escapes such as \d and \s have NOT been changed.
1830 16. Updated pcredemo.c to include calls to free() to release the memory for the
1831 compiled pattern.
1833 17. The generated file chartables.c was being created in the source directory
1834 instead of in the building directory. This caused the build to fail if the
1835 source directory was different from the building directory, and was
1836 read-only.
1838 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1839 file. No doubt somebody will tell me if they don't make sense... Also added
1840 Dan Mooney's comments about building on OpenVMS.
1842 19. Added support for partial matching via the PCRE_PARTIAL option for
1843 pcre_exec() and the \P data escape in pcretest.
1845 20. Extended pcretest with 3 new pattern features:
1847 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1848 write the compiled pattern to the file whose name is "rest-of-line".
1849 This is a straight binary dump of the data, with the saved pointer to
1850 the character tables forced to be NULL. The study data, if any, is
1851 written too. After writing, pcretest reads a new pattern.
1853 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1854 compiled pattern from the given file. There must not be any
1855 occurrences of "<" in the file name (pretty unlikely); if there are,
1856 pcretest will instead treat the initial "<" as a pattern delimiter.
1857 After reading in the pattern, pcretest goes on to read data lines as
1858 usual.
1860 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1861 and 16-bit fields in a compiled pattern, to simulate a pattern that
1862 was compiled on a host of opposite endianness.
1864 21. The pcre-exec() function can now cope with patterns that were compiled on
1865 hosts of opposite endianness, with this restriction:
1867 As for any compiled expression that is saved and used later, the tables
1868 pointer field cannot be preserved; the extra_data field in the arguments
1869 to pcre_exec() should be used to pass in a tables address if a value
1870 other than the default internal tables were used at compile time.
1872 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1873 now diagnosed as an error. Previously, most of the time, a negative number
1874 would have been treated as zero, but if in addition "ovector" was passed as
1875 NULL, a crash could occur.
1877 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1878 new versions from the libtool 1.5 distribution (the last one is a copy of
1879 a file called libtool.m4). This seems to have fixed the need to patch
1880 "configure" to support Darwin 1.3 (which I used to do). However, I still
1881 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1882 workstation).
1884 24. Changed the PCRE licence to be the more standard "BSD" licence.
1887 Version 4.5 01-Dec-03
1888 ---------------------
1890 1. There has been some re-arrangement of the code for the match() function so
1891 that it can be compiled in a version that does not call itself recursively.
1892 Instead, it keeps those local variables that need separate instances for
1893 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1894 needs to "recurse". Keeping track of where control must go is done by means
1895 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1896 hide most of the details from the main code, and operates only if
1897 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1898 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1899 operating.
1901 To make it easier for callers to provide specially tailored get/free
1902 functions for this usage, two new functions, pcre_stack_malloc, and
1903 pcre_stack_free, are used. They are always called in strict stacking order,
1904 and the size of block requested is always the same.
1906 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1907 PCRE has been compiled to use the stack or the heap for recursion. The
1908 -C option of pcretest uses this to show which version is compiled.
1910 A new data escape \S, is added to pcretest; it causes the amounts of store
1911 obtained and freed by both kinds of malloc/free at match time to be added
1912 to the output.
1914 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1915 what's available on my current Linux desktop machine.
1917 3. When matching a UTF-8 string, the test for a valid string at the start has
1918 been extended. If start_offset is not zero, PCRE now checks that it points
1919 to a byte that is the start of a UTF-8 character. If not, it returns
1920 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1921 this is necessary because there may be backward assertions in the pattern.
1922 When matching the same subject several times, it may save resources to use
1923 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1925 4. The code for checking the validity of UTF-8 strings has been tightened so
1926 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1927 containing "overlong sequences".
1929 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1930 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1931 should have been "&", but it just so happened that all the cases this let
1932 through by mistake were picked up later in the function.
1934 6. I had used a variable called "isblank" - this is a C99 function, causing
1935 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1937 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1938 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1939 so the version is shown; (c) stop "make test" echoing "./RunTest".
1941 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1943 9. The prototype for memmove() for systems that don't have it was using
1944 size_t, but the inclusion of the header that defines size_t was later. I've
1945 moved the #includes for the C headers earlier to avoid this.
1947 10. Added some adjustments to the code to make it easier to compiler on certain
1948 special systems:
1950 (a) Some "const" qualifiers were missing.
1951 (b) Added the macro EXPORT before all exported functions; by default this
1952 is defined to be empty.
1953 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1954 that it reads its output file name as an argument instead of writing
1955 to the standard output and assuming this can be redirected.
1957 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1958 class containing characters with values greater than 255, PCRE compilation
1959 went into a loop.
1961 12. A recursive reference to a subpattern that was within another subpattern
1962 that had a minimum quantifier of zero caused PCRE to crash. For example,
1963 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1964 recursion. If the recursively-called subpattern itself had a zero repeat,
1965 that was OK.
1967 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1968 buffer into which it was copied (for escape processing) was still set at
1969 1024, so long lines caused crashes.
1971 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1972 "internal error: code overflow...". This applied to any character class
1973 that was followed by a possessive quantifier.
1975 15. Modified the Makefile to add libpcre.la as a prerequisite for
1976 libpcreposix.la because I was told this is needed for a parallel build to
1977 work.
1979 16. If a pattern that contained .* following optional items at the start was
1980 studied, the wrong optimizing data was generated, leading to matching
1981 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1982 matching string must start with a or b or c. The correct conclusion for
1983 this pattern is that a match can start with any character.
1986 Version 4.4 13-Aug-03
1987 ---------------------
1989 1. In UTF-8 mode, a character class containing characters with values between
1990 127 and 255 was not handled correctly if the compiled pattern was studied.
1991 In fixing this, I have also improved the studying algorithm for such
1992 classes (slightly).
1994 2. Three internal functions had redundant arguments passed to them. Removal
1995 might give a very teeny performance improvement.
1997 3. Documentation bug: the value of the capture_top field in a callout is *one
1998 more than* the number of the hightest numbered captured substring.
2000 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2001 in incorrectly linking with a previously installed version. They now link
2002 explicitly with libpcre.la.
2004 5. configure.in no longer needs to recognize Cygwin specially.
2006 6. A problem in pcre.in for Windows platforms is fixed.
2008 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2009 pcretest, it used to include the size of the study block as part of its
2010 output. Unfortunately, the structure contains a field that has a different
2011 size on different hardware architectures. This meant that the tests that
2012 showed this size failed. As the block is currently always of a fixed size,
2013 this information isn't actually particularly useful in pcretest output, so
2014 I have just removed it.
2016 8. Three pre-processor statements accidentally did not start in column 1.
2017 Sadly, there are *still* compilers around that complain, even though
2018 standard C has not required this for well over a decade. Sigh.
2020 9. In pcretest, the code for checking callouts passed small integers in the
2021 callout_data field, which is a void * field. However, some picky compilers
2022 complained about the casts involved for this on 64-bit systems. Now
2023 pcretest passes the address of the small integer instead, which should get
2024 rid of the warnings.
2026 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2027 both compile and run time, and gives an error if an invalid UTF-8 sequence
2028 is found. There is a option for disabling this check in cases where the
2029 string is known to be correct and/or the maximum performance is wanted.
2031 11. In response to a bug report, I changed one line in Makefile.in from
2033 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2034 to
2035 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2037 to look similar to other lines, but I have no way of telling whether this
2038 is the right thing to do, as I do not use Windows. No doubt I'll get told
2039 if it's wrong...
2042 Version 4.3 21-May-03
2043 ---------------------
2045 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2046 Makefile.
2048 2. Some refactoring to improve the quality of the code:
2050 (i) The utf8_table... variables are now declared "const".
2052 (ii) The code for \cx, which used the "case flipping" table to upper case
2053 lower case letters, now just substracts 32. This is ASCII-specific,
2054 but the whole concept of \cx is ASCII-specific, so it seems
2055 reasonable.
2057 (iii) PCRE was using its character types table to recognize decimal and
2058 hexadecimal digits in the pattern. This is silly, because it handles
2059 only 0-9, a-f, and A-F, but the character types table is locale-
2060 specific, which means strange things might happen. A private
2061 table is now used for this - though it costs 256 bytes, a table is
2062 much faster than multiple explicit tests. Of course, the standard
2063 character types table is still used for matching digits in subject
2064 strings against \d.
2066 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2067 ending in _t are). So I've renamed it as ESC_tee.
2069 3. The first argument for regexec() in the POSIX wrapper should have been
2070 defined as "const".
2072 4. Changed pcretest to use malloc() for its buffers so that they can be
2073 Electric Fenced for debugging.
2075 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2076 to read one or more bytes before the start of the subject string. Often this
2077 had no effect on PCRE's behaviour, but in some circumstances it could
2078 provoke a segmentation fault.
2080 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2081 to try to read one or more bytes before the start of the subject string.
2083 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2084 UTF-8 support could misbehave in various ways if the subject string
2085 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2086 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2087 back over UTF-8 characters.)
2090 Version 4.2 14-Apr-03
2091 ---------------------
2093 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2095 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2096 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2097 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2098 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2099 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2101 Note: automatic setting of the BUILD variables is not yet working
2102 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2103 compile-time but not at link-time
2104 [LINK]: use for linking executables only
2105 make different versions for Windows and non-Windows
2106 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2107 libraries
2108 [LINK_FOR_BUILD]: new variable
2109 [OBJEXT]: use throughout
2110 [EXEEXT]: use throughout
2111 <winshared>: new target
2112 <wininstall>: new target
2113 <dftables.o>: use native compiler
2114 <dftables>: use native linker
2115 <install>: handle Windows platform correctly
2116 <clean>: ditto
2117 <check>: ditto
2118 copy DLL to top builddir before testing
2120 As part of these changes, -no-undefined was removed again. This was reported
2121 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2122 in any case.
2124 3. Some tidies to get rid of compiler warnings:
2126 . In the match_data structure, match_limit was an unsigned long int, whereas
2127 match_call_count was an int. I've made them both unsigned long ints.
2129 . In pcretest the fact that a const uschar * doesn't automatically cast to
2130 a void * provoked a warning.
2132 . Turning on some more compiler warnings threw up some "shadow" variables
2133 and a few more missing casts.
2135 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2136 option, a class that contained a single character with a value between 128
2137 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2139 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2140 option, a class that contained several characters, but with at least one
2141 whose value was between 128 and 255 caused PCRE to crash.
2144 Version 4.1 12-Mar-03
2145 ---------------------
2147 1. Compiling with gcc -pedantic found a couple of places where casts were
2148 needed, and a string in dftables.c that was longer than standard compilers are
2149 required to support.
2151 2. Compiling with Sun's compiler found a few more places where the code could
2152 be tidied up in order to avoid warnings.
2154 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2155 first of these names is deprecated in the latest Autoconf in favour of the name
2156 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2157 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2158 analogy I have changed it to CFLAGS_FOR_BUILD.
2160 4. Added -no-undefined to the linking command in the Makefile, because this is
2161 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2162 linking step for the pcreposix library.
2164 5. PCRE was failing to diagnose the case of two named groups with the same
2165 name.
2167 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2168 literal character that is needed in the subject for a match, and scans along to
2169 ensure that it is present before embarking on the full matching process. This
2170 saves time in cases of nested unlimited repeats that are never going to match.
2171 Problem: the scan can take a lot of time if the subject is very long (e.g.
2172 megabytes), thus penalizing straightforward matches. It is now done only if the
2173 amount of subject to be scanned is less than 1000 bytes.
2175 7. A lesser problem with the same optimization is that it was recording the
2176 first character of an anchored pattern as "needed", thus provoking a search
2177 right along the subject, even when the first match of the pattern was going to
2178 fail. The "needed" character is now not set for anchored patterns, unless it
2179 follows something in the pattern that is of non-fixed length. Thus, it still
2180 fulfils its original purpose of finding quick non-matches in cases of nested
2181 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2184 Version 4.0 17-Feb-03
2185 ---------------------
2187 1. If a comment in an extended regex that started immediately after a meta-item
2188 extended to the end of string, PCRE compiled incorrect data. This could lead to
2189 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2191 2. Moved to autoconf 2.53 and libtool 1.4.2.
2193 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2194 the special perltest8 script is no longer needed - all the tests can be run
2195 from a single perltest script.
2197 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2198 by \s. It has now been removed in PCRE. This means it isn't recognized as
2199 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2200 class [:space:] *does* include VT, thereby creating a mess.
2202 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2203 space and tab.
2205 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2206 its new features into the main test script, reducing the number of scripts.
2208 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2209 were backward compatible, and made the (?i) apply to the whole pattern, as if
2210 /i were given. Now it behaves more logically, and applies the option setting
2211 only to what follows. PCRE has been changed to follow suit. However, if it
2212 finds options settings right at the start of the pattern, it extracts them into
2213 the global options, as before. Thus, they show up in the info data.
2215 8. Added support for the \Q...\E escape sequence. Characters in between are
2216 treated as literals. This is slightly different from Perl in that $ and @ are
2217 also handled as literals inside the quotes. In Perl, they will cause variable
2218 interpolation. Note the following examples:
2220 Pattern PCRE matches Perl matches
2222 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2223 \Qabc\$xyz\E abc\$xyz abc\$xyz
2224 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2226 For compatibility with Perl, \Q...\E sequences are recognized inside character
2227 classes as well as outside them.
2229 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2230 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2231 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2232 signed/unsigned warnings.
2234 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2235 option for pcretest, so I've replaced it by a simple function that does just
2236 that job.
2238 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2239 "pcregrep -".
2241 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2242 Java package. This provides some syntactic sugar for simple cases of what my
2243 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2244 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2245 item, you can use this simplified notation. Note that only makes sense with
2246 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2247 greediness, whatever the setting of the PCRE_UNGREEDY option.
2249 13. A change of greediness default within a pattern was not taking effect at
2250 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2251 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2252 was abstracted outside.
2254 14. PCRE now supports the \G assertion. It is true when the current matching
2255 position is at the start point of the match. This differs from \A when the
2256 starting offset is non-zero. Used with the /g option of pcretest (or similar
2257 code), it works in the same way as it does for Perl's /g option. If all
2258 alternatives of a regex begin with \G, the expression is anchored to the start
2259 match position, and the "anchored" flag is set in the compiled expression.
2261 15. Some bugs concerning the handling of certain option changes within patterns
2262 have been fixed. These applied to options other than (?ims). For example,
2263 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2264 been the other way round. Some of this was related to change 7 above.
2266 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2267 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2268 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2269 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2271 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2272 mode. Unlike ".", it always matches newline, whatever the setting of
2273 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2274 assertions. Perl allows it, but it doesn't (in general) work because it can't
2275 calculate the length of the lookbehind. At least, that's the case for Perl
2276 5.8.0 - I've been told they are going to document that it doesn't work in
2277 future.
2279 18. Added an error diagnosis for escapes that PCRE does not support: these are
2280 \L, \l, \N, \P, \p, \U, \u, and \X.
2282 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2283 reading past the end of the pattern in cases such as /[abcd/.
2285 20. PCRE was getting more memory than necessary for patterns with classes that
2286 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2288 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2289 compiling PCRE for use with Virtual Pascal.
2291 22. Small fix to the Makefile to make it work properly if the build is done
2292 outside the source tree.
2294 23. Added a new extension: a condition to go with recursion. If a conditional
2295 subpattern starts with (?(R) the "true" branch is used if recursion has
2296 happened, whereas the "false" branch is used only at the top level.
2298 24. When there was a very long string of literal characters (over 255 bytes
2299 without UTF support, over 250 bytes with UTF support), the computation of how
2300 much memory was required could be incorrect, leading to segfaults or other
2301 strange effects.
2303 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2304 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2305 there was a subsequent back reference to those brackets. This meant that, for
2306 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2307 possible to check for precisely this case. All we can do is abandon the
2308 optimization if .* occurs inside capturing brackets when there are any back
2309 references whatsoever. (See below for a better fix that came later.)
2311 26. The handling of the optimization for finding the first character of a
2312 non-anchored pattern, and for finding a character that is required later in the
2313 match were failing in some cases. This didn't break the matching; it just
2314 failed to optimize when it could. The way this is done has been re-implemented.
2316 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2318 28. Added a new feature that provides some of the functionality that Perl
2319 provides with (?{...}). The facility is termed a "callout". The way it is done
2320 in PCRE is for the caller to provide an optional function, by setting
2321 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2322 global variable. By default it is unset, which disables all calling out. To get
2323 the function called, the regex must include (?C) at appropriate points. This
2324 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2325 This provides a means of identifying different callout points. When PCRE
2326 reaches such a point in the regex, if pcre_callout has been set, the external
2327 function is called. It is provided with data in a structure called
2328 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2329 matching continues; if it returns a non-zero value, the match at the current
2330 point fails. However, backtracking will occur if possible. [This was changed
2331 later and other features added - see item 49 below.]
2333 29. pcretest is upgraded to test the callout functionality. It provides a
2334 callout function that displays information. By default, it shows the start of
2335 the match and the current position in the text. There are some new data escapes
2336 to vary what happens:
2338 \C+ in addition, show current contents of captured substrings
2339 \C- do not supply a callout function
2340 \C!n return 1 when callout number n is reached
2341 \C!n!m return 1 when callout number n is reached for the mth time
2343 30. If pcregrep was called with the -l option and just a single file name, it
2344 output "<stdin>" if a match was found, instead of the file name.
2346 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2347 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2348 pcre_exec(). This saves a malloc/free per call. The default value of
2349 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2350 when configuring.
2352 32. The default maximum size of a compiled pattern is 64K. There have been a
2353 few cases of people hitting this limit. The code now uses macros to handle the
2354 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2355 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2356 configuring. Tests 2 and 5 work only with 2-byte links because they output
2357 debugging information about compiled patterns.
2359 33. Internal code re-arrangements:
2361 (a) Moved the debugging function for printing out a compiled regex into
2362 its own source file (printint.c) and used #include to pull it into
2363 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2364 separate copies.
2366 (b) Defined the list of op-code names for debugging as a macro in
2367 internal.h so that it is next to the definition of the opcodes.
2369 (c) Defined a table of op-code lengths for simpler skipping along compiled
2370 code. This is again a macro in internal.h so that it is next to the
2371 definition of the opcodes.
2373 34. Added support for recursive calls to individual subpatterns, along the
2374 lines of Robin Houston's patch (but implemented somewhat differently).
2376 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2377 allow it to read and process whole directories in Win32. This code was
2378 contributed by Lionel Fourquaux; it has not been tested by me.
2380 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2381 used to name a group. Names consist of alphanumerics and underscores, and must
2382 be unique. Back references use the syntax (?P=name) and recursive calls use
2383 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2384 numbers. The function pcre_fullinfo() can be used after compilation to extract
2385 a name/number map. There are three relevant calls:
2387 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2388 PCRE_INFO_NAMECOUNT yields the number of entries
2389 PCRE_INFO_NAMETABLE yields a pointer to the map.
2391 The map is a vector of fixed-size entries. The size of each entry depends on
2392 the length of the longest name used. The first two bytes of each entry are the
2393 group number, most significant byte first. There follows the corresponding
2394 name, zero terminated. The names are in alphabetical order.
2396 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2397 case instead of 255. Making it the same both with and without UTF-8 support
2398 means that the same test output works with both.
2400 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2401 calling malloc() with a zero argument.
2403 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2404 optimization. I've improved things by keeping a bitmap of backreferences with
2405 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2406 fact referenced, the optimization can be applied. It is unlikely that a
2407 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2408 the match to follow \n) will appear inside brackets with a number greater than
2409 31, but if it does, any back reference > 31 suppresses the optimization.
2411 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2412 of disabling numbered capturing parentheses. Any opening parenthesis that is
2413 not followed by ? behaves as if it were followed by ?: but named parentheses
2414 can still be used for capturing (and they will acquire numbers in the usual
2415 way).
2417 41. Redesigned the return codes from the match() function into yes/no/error so
2418 that errors can be passed back from deep inside the nested calls. A malloc
2419 failure while inside a recursive subpattern call now causes the
2420 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2422 42. It is now possible to set a limit on the number of times the match()
2423 function is called in a call to pcre_exec(). This facility makes it possible to
2424 limit the amount of recursion and backtracking, though not in a directly
2425 obvious way, because the match() function is used in a number of different
2426 circumstances. The count starts from zero for each position in the subject
2427 string (for non-anchored patterns). The default limit is, for compatibility, a
2428 large number, namely 10 000 000. You can change this in two ways:
2430 (a) When configuring PCRE before making, you can use --with-match-limit=n
2431 to set a default value for the compiled library.
2433 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2434 a different value is set. See 45 below.
2436 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2438 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2439 of things that can be changed at compile time. The first argument specifies
2440 what is wanted and the second points to where the information is to be placed.
2441 The current list of available information is:
2445 The output is an integer that is set to one if UTF-8 support is available;
2446 otherwise it is set to zero.
2450 The output is an integer that it set to the value of the code that is used for
2451 newline. It is either LF (10) or CR (13).
2455 The output is an integer that contains the number of bytes used for internal
2456 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2460 The output is an integer that contains the threshold above which the POSIX
2461 interface uses malloc() for output vectors. See item 31 above.
2465 The output is an unsigned integer that contains the default limit of the number
2466 of match() calls in a pcre_exec() execution. See 42 above.
2468 44. pcretest has been upgraded by the addition of the -C option. This causes it
2469 to extract all the available output from the new pcre_config() function, and to
2470 output it. The program then exits immediately.
2472 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2473 order to support additional features. One way would have been to define
2474 pcre_exec2() (for example) with extra arguments, but this would not have been
2475 extensible, and would also have required all calls to the original function to
2476 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2477 is used for passing in "extra" data from pcre_study().
2479 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2480 contains the following fields:
2482 flags a bitmap indicating which of the following fields are set
2483 study_data opaque data from pcre_study()
2484 match_limit a way of specifying a limit on match() calls for a specific
2485 call to pcre_exec()
2486 callout_data data for callouts (see 49 below)
2488 The flag bits are also defined in pcre.h, and are
2494 The pcre_study() function now returns one of these new pcre_extra blocks, with
2495 the actual study data pointed to by the study_data field, and the
2496 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2497 before. That is, this change is entirely upwards-compatible and requires no
2498 change to existing code.
2500 If you want to pass in additional data to pcre_exec(), you can either place it
2501 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2502 block.
2504 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2505 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2506 times with different match limits, until it finds the minimum value needed for
2507 pcre_exec() to complete. The value is then output. This can be instructive; for
2508 most simple matches the number is quite small, but for pathological cases it
2509 gets very large very quickly.
2511 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2512 returns the size of the data block pointed to by the study_data field in a
2513 pcre_extra block, that is, the value that was passed as the argument to
2514 pcre_malloc() when PCRE was getting memory in which to place the information
2515 created by pcre_study(). The fourth argument should point to a size_t variable.
2516 pcretest has been extended so that this information is shown after a successful
2517 pcre_study() call when information about the compiled regex is being displayed.
2519 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2520 because what follows is always an absolute path. (Later: it turns out that this
2521 is more than cosmetic for MinGW, because it doesn't like empty path
2522 components.)
2524 49. Some changes have been made to the callout feature (see 28 above):
2526 (i) A callout function now has three choices for what it returns:
2528 0 => success, carry on matching
2529 > 0 => failure at this point, but backtrack if possible
2530 < 0 => serious error, return this value from pcre_exec()
2532 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2533 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2534 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2535 use by callout functions. It will never be used by PCRE itself.
2537 (ii) The pcre_extra structure (see 45 above) has a void * field called
2538 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2539 pcre_callout_block structure has a field of the same name. The contents of
2540 the field passed in the pcre_extra structure are passed to the callout
2541 function in the corresponding field in the callout block. This makes it
2542 easier to use the same callout-containing regex from multiple threads. For
2543 testing, the pcretest program has a new data escape
2545 \C*n pass the number n (may be negative) as callout_data
2547 If the callout function in pcretest receives a non-zero value as
2548 callout_data, it returns that value.
2550 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2551 there were some redundant $(CFLAGS) in commands that are now specified as
2552 $(LINK), which already includes $(CFLAGS).
2554 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2555 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2556 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2557 one-byte characters throughout. Note that case-insensitive matching applies
2558 only to characters whose values are less than 256. PCRE doesn't support the
2559 notion of cases for higher-valued characters.
2561 (i) A character class whose characters are all within 0-255 is handled as
2562 a bit map, and the map is inverted for negative classes. Previously, a
2563 character > 255 always failed to match such a class; however it should
2564 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2566 (ii) A negated character class with a single character < 255 is coded as
2567 "not this character" (OP_NOT). This wasn't working properly when the test
2568 character was multibyte, either singly or repeated.
2570 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2571 mode, for example: \x{100}{2,3}.
2573 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2574 singly or repeated) now correctly test multibyte characters. However,
2575 PCRE doesn't recognize any characters with values greater than 255 as
2576 digits, spaces, or word characters. Such characters always match \D, \S,
2577 and \W, and never match \d, \s, or \w.
2579 (v) Classes may now contain characters and character ranges with values
2580 greater than 255. For example: [ab\x{100}-\x{400}].
2582 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2583 PCRE in UTF-8 mode.
2585 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2586 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2587 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2588 value.)
2590 53. The single man page has become too large. I have therefore split it up into
2591 a number of separate man pages. These also give rise to individual HTML pages;
2592 these are now put in a separate directory, and there is an index.html page that
2593 lists them all. Some hyperlinking between the pages has been installed.
2595 54. Added convenience functions for handling named capturing parentheses.
2597 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2598 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2599 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2600 are faulted.
2602 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2603 calling configure. These values are used when compiling the dftables.c program
2604 which is run to generate the source of the default character tables. They
2605 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2606 you will need to set these values.
2608 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2611 Version 3.9 02-Jan-02
2612 ---------------------
2614 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2616 2. If --disable-static was given, the building process failed when trying to
2617 build pcretest and pcregrep. (For some reason it was using libtool to compile
2618 them, which is not right, as they aren't part of the library.)
2621 Version 3.8 18-Dec-01
2622 ---------------------
2624 1. The experimental UTF-8 code was completely screwed up. It was packing the
2625 bytes in the wrong order. How dumb can you get?
2628 Version 3.7 29-Oct-01
2629 ---------------------
2631 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2632 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2633 this didn't happen under Solaris 8, where I normally test things.
2635 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2636 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2637 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2638 this fix an hour or so after the initial 3.7 release.)
2641 Version 3.6 23-Oct-01
2642 ---------------------
2644 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2645 offsets passed as NULL with zero offset count.
2647 2. The config.guess and config.sub files had not been updated when I moved to
2648 the latest autoconf.
2651 Version 3.5 15-Aug-01
2652 ---------------------
2654 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2655 had been forgotten.
2657 2. By using declared but undefined structures, we can avoid using "void"
2658 definitions in pcre.h while keeping the internal definitions of the structures
2659 private.
2661 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2662 user point of view, this means that both static and shared libraries are built
2663 by default, but this can be individually controlled. More of the work of
2664 handling this static/shared cases is now inside libtool instead of PCRE's make
2665 file.
2667 4. The pcretest utility is now installed along with pcregrep because it is
2668 useful for users (to test regexs) and by doing this, it automatically gets
2669 relinked by libtool. The documentation has been turned into a man page, so
2670 there are now .1, .txt, and .html versions in /doc.
2672 5. Upgrades to pcregrep:
2673 (i) Added long-form option names like gnu grep.
2674 (ii) Added --help to list all options with an explanatory phrase.
2675 (iii) Added -r, --recursive to recurse into sub-directories.
2676 (iv) Added -f, --file to read patterns from a file.
2678 6. pcre_exec() was referring to its "code" argument before testing that
2679 argument for NULL (and giving an error if it was NULL).
2681 7. Upgraded Makefile.in to allow for compiling in a different directory from
2682 the source directory.
2684 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2685 options bits, the pointer it was passed was to an int instead of to an unsigned
2686 long int. This mattered only on 64-bit systems.
2688 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2689 generated) instead of pcre.in, which it its source. Also made the same change
2690 in several of the .c files.
2692 10. A new release of gcc defines printf() as a macro, which broke pcretest
2693 because it had an ifdef in the middle of a string argument for printf(). Fixed
2694 by using separate calls to printf().
2696 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2697 script, to force use of CR or LF instead of \n in the source. On non-Unix
2698 systems, the value can be set in config.h.
2700 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2701 absolute limit. Changed the text of the error message to make this clear, and
2702 likewise updated the man page.
2704 13. The limit of 99 on the number of capturing subpatterns has been removed.
2705 The new limit is 65535, which I hope will not be a "real" limit.
2708 Version 3.4 22-Aug-00
2709 ---------------------
2711 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2713 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2716 Version 3.3 01-Aug-00
2717 ---------------------
2719 1. If an octal character was given, but the value was greater than \377, it
2720 was not getting masked to the least significant bits, as documented. This could
2721 lead to crashes in some systems.
2723 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2724 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2726 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2727 These just pass their arguments on to (pcre_free)(), but they are provided
2728 because some uses of PCRE bind it to non-C systems that can call its functions,
2729 but cannot call free() or pcre_free() directly.
2731 4. Add "make test" as a synonym for "make check". Corrected some comments in
2732 the Makefile.
2734 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2735 Makefile.
2737 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2738 command called pgrep for grepping around the active processes.
2740 7. Added the beginnings of support for UTF-8 character strings.
2742 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2743 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2744 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2745 out for the ar command.)
2748 Version 3.2 12-May-00
2749 ---------------------
2751 This is purely a bug fixing release.
2753 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2754 of ZA. This was just one example of several cases that could provoke this bug,
2755 which was introduced by change 9 of version 2.00. The code for breaking
2756 infinite loops after an iteration that matches an empty string was't working
2757 correctly.
2759 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2760 when matched against abbab (for example). After matching an empty string, it
2761 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2762 caused it to match further down the string than it should.
2764 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2765 was there because it doesn't seem to be needed, and it causes trouble on some
2766 systems, as it is not a Standard C header. It has been removed.
2768 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2769 were reported on the Macintosh. The changes were from
2771 while ((c = *(++ptr)) != 0 && c != '\n');
2772 to
2773 while ((c = *(++ptr)) != 0 && c != '\n') ;
2775 Totally extraordinary, but if that's what it takes...
2777 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2778 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2779 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2780 assumes the way PCRE uses memmove() (always moving upwards).
2782 6. PCRE is being used in one environment where strchr() is not available. There
2783 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2784 faster code anyway.
2787 Version 3.1 09-Feb-00
2788 ---------------------
2790 The only change in this release is the fixing of some bugs in Makefile.in for
2791 the "install" target:
2793 (1) It was failing to install pcreposix.h.
2795 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2798 Version 3.0 01-Feb-00
2799 ---------------------
2801 1. Add support for the /+ modifier to perltest (to output $` like it does in
2802 pcretest).
2804 2. Add support for the /g modifier to perltest.
2806 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2807 matches null strings.
2809 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2810 pattern. Perl treats empty patterns specially - it reuses the most recent
2811 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2812 effect.
2814 5. The POSIX interface was broken in that it was just handing over the POSIX
2815 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2816 required a bigger vector, with some working space on the end. This means that
2817 the POSIX wrapper now has to get and free some memory, and copy the results.
2819 6. Added some simple autoconf support, placing the test data and the
2820 documentation in separate directories, re-organizing some of the
2821 information files, and making it build pcre-config (a GNU standard). Also added
2822 libtool support for building PCRE as a shared library, which is now the
2823 default.
2825 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2826 09 are not valid octal constants. Single digits will be used for minor values
2827 less than 10.
2829 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2830 existing programs that set these in the POSIX interface can use PCRE without
2831 modification.
2833 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2834 return all that pcre_info() returns, plus additional data. The pcre_info()
2835 function is retained for compatibility, but is considered to be obsolete.
2837 10. Added experimental recursion feature (?R) to handle one common case that
2838 Perl 5.6 will be able to do with (?p{...}).
2840 11. Added support for POSIX character classes like [:alpha:], which Perl is
2841 adopting.
2844 Version 2.08 31-Aug-99
2845 ----------------------
2847 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2848 trying to match at the startoffset position, but instead was moving forward to
2849 the next newline as if a previous match had failed.
2851 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2852 and could get into a loop if a null string was matched other than at the start
2853 of the subject.
2855 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2856 be distinguished at compile time, and for completeness also added PCRE_DATE.
2858 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2859 in GnuWin32 environments.
2862 Version 2.07 29-Jul-99
2863 ----------------------
2865 1. The documentation is now supplied in plain text form and HTML as well as in
2866 the form of man page sources.
2868 2. C++ compilers don't like assigning (void *) values to other pointer types.
2869 In particular this affects malloc(). Although there is no problem in Standard
2870 C, I've put in casts to keep C++ compilers happy.
2872 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2873 should be (const char *).
2875 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2876 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2877 However, I haven't made this a standard facility. The documentation doesn't
2878 mention it, and the Makefile doesn't support it.
2880 5. The Makefile now contains an "install" target, with editable destinations at
2881 the top of the file. The pcretest program is not installed.
2883 6. pgrep -V now gives the PCRE version number and date.
2885 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2886 causing the entire string to be ignored, instead of just the last character.
2888 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2889 non-matching string, it can take a very, very long time, even for strings of
2890 quite modest length, because of the nested recursion. PCRE now does better in
2891 some of these cases. It does this by remembering the last required literal
2892 character in the pattern, and pre-searching the subject to ensure it is present
2893 before running the real match. In other words, it applies a heuristic to detect
2894 some types of certain failure quickly, and in the above example, if presented
2895 with a string that has no trailing " it gives "no match" very quickly.
2897 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2898 other alternatives are tried instead.
2901 Version 2.06 09-Jun-99
2902 ----------------------
2904 1. Change pcretest's output for amount of store used to show just the code
2905 space, because the remainder (the data block) varies in size between 32-bit and
2906 64-bit systems.
2908 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2909 start matching at. This allows lookbehinds to work when searching for multiple
2910 occurrences in a string.
2912 3. Added additional options to pcretest for testing multiple occurrences:
2914 /+ outputs the rest of the string that follows a match
2915 /g loops for multiple occurrences, using the new startoffset argument
2916 /G loops for multiple occurrences by passing an incremented pointer
2918 4. PCRE wasn't doing the "first character" optimization for patterns starting
2919 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2920 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2921 the letter 'x'. On long subject strings, this gives a significant speed-up.
2924 Version 2.05 21-Apr-99
2925 ----------------------
2927 1. Changed the type of magic_number from int to long int so that it works
2928 properly on 16-bit systems.
2930 2. Fixed a bug which caused patterns starting with .* not to work correctly
2931 when the subject string contained newline characters. PCRE was assuming
2932 anchoring for such patterns in all cases, which is not correct because .* will
2933 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2934 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2935 must be retried after every newline in the subject.
2938 Version 2.04 18-Feb-99
2939 ----------------------
2941 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2942 computation of the store needed to hold the pattern was incorrect (too large).
2943 If such patterns were nested a few deep, this could multiply and become a real
2944 problem.
2946 2. Added /M option to pcretest to show the memory requirement of a specific
2947 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2949 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2950 compiled in such a way that the backtracking after subsequent failure was
2951 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2952 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2955 Version 2.03 02-Feb-99
2956 ----------------------
2958 1. Fixed typo and small mistake in man page.
2960 2. Added 4th condition (GPL supersedes if conflict) and created separate
2961 LICENCE file containing the conditions.
2963 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2964 Perl, that is the internal \ allows the delimiter to be included in the
2965 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2966 the final delimiter, add \ to the end of the pattern (to test the error).
2968 4. Added the convenience functions for extracting substrings after a successful
2969 match. Updated pcretest to make it able to test these functions.
2972 Version 2.02 14-Jan-99
2973 ----------------------
2975 1. Initialized the working variables associated with each extraction so that
2976 their saving and restoring doesn't refer to uninitialized store.
2978 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2979 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2980 fix the problem.
2982 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2983 calls, and wasn't printing the correct value for compiling calls. Increased the
2984 default value of LOOPREPEAT, and the number of significant figures in the
2985 times.
2987 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2989 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2990 a building problem on Windows NT with a FAT file system.
2993 Version 2.01 21-Oct-98
2994 ----------------------
2996 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2997 to character tables built by pcre_maketables() in the current locale. If NULL
2998 is passed, the default tables are used.
3001 Version 2.00 24-Sep-98
3002 ----------------------
3004 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3005 it any more.
3007 2. Allow quantification of (?>) groups, and make it work correctly.
3009 3. The first character computation wasn't working for (?>) groups.
3011 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3012 end of the subject) and add 5.005's \z, which really does match only at the
3013 very end of the subject.
3015 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3017 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3018 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3019 localized options. All options to pcre_study() were also removed.
3021 7. Add other new features from 5.005:
3023 $(?<= positive lookbehind
3024 $(?<! negative lookbehind
3025 (?imsx-imsx) added the unsetting capability
3026 such a setting is global if at outer level; local otherwise
3027 (?imsx-imsx:) non-capturing groups with option setting
3028 (?(cond)re|re) conditional pattern matching
3030 A backreference to itself in a repeated group matches the previous
3031 captured string.
3033 8. General tidying up of studying (both automatic and via "study")
3034 consequential on the addition of new assertions.
3036 9. As in 5.005, unlimited repeated groups that could match an empty substring
3037 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3038 runtime if any iteration does actually match an empty substring.
3040 10. Include the RunTest script in the distribution.
3042 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3043 discrepancies, some of which were old and were also with respect to 5.004. They
3044 have now been fixed.
3047 Version 1.09 28-Apr-98
3048 ----------------------
3050 1. A negated single character class followed by a quantifier with a minimum
3051 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3052 program crashes, or just wrong answers. This did not apply to negated classes
3053 containing more than one character, or to minima other than one.
3056 Version 1.08 27-Mar-98
3057 ----------------------
3059 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3061 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3062 latter must appear before anything that relies on it in the pattern.
3065 Version 1.07 16-Feb-98
3066 ----------------------
3068 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3069 repeat of a potentially empty string).
3072 Version 1.06 23-Jan-98
3073 ----------------------
3075 1. Added Markus Oberhumer's little patches for C++.
3077 2. Literal strings longer than 255 characters were broken.
3080 Version 1.05 23-Dec-97
3081 ----------------------
3083 1. Negated character classes containing more than one character were failing if
3084 PCRE_CASELESS was set at run time.
3087 Version 1.04 19-Dec-97
3088 ----------------------
3090 1. Corrected the man page, where some "const" qualifiers had been omitted.
3092 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3093 input syntax.
3095 3. Fixed memory leak which occurred when a regex with back references was
3096 matched with an offsets vector that wasn't big enough. The temporary memory
3097 that is used in this case wasn't being freed if the match failed.
3099 4. Tidied pcretest to ensure it frees memory that it gets.
3101 5. Temporary memory was being obtained in the case where the passed offsets
3102 vector was exactly big enough.
3104 6. Corrected definition of offsetof() from change 5 below.
3106 7. I had screwed up change 6 below and broken the rules for the use of
3107 setjmp(). Now fixed.
3110 Version 1.03 18-Dec-97
3111 ----------------------
3113 1. A erroneous regex with a missing opening parenthesis was correctly
3114 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3115 on some systems.
3117 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3118 it was reported that one broken compiler failed on the former because "code" is
3119 also an independent variable.
3121 3. The erroneous regex a[]b caused an array overrun reference.
3123 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3124 fail on data ending with that character. (It was going on too far, and checking
3125 the next character, typically a binary zero.) This was specific to the
3126 optimized code for single-character negative classes.
3128 5. Added a contributed patch from the TIN world which does the following:
3130 + Add an undef for memmove, in case the the system defines a macro for it.
3132 + Add a definition of offsetof(), in case there isn't one. (I don't know
3133 the reason behind this - offsetof() is part of the ANSI standard - but
3134 it does no harm).
3136 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3137 most of the places where whitespace preceded '#'. I have given up and
3138 allowed the remaining 2 cases to be at the margin.
3140 + Rename some variables in pcre to eliminate shadowing. This seems very
3141 pedantic, but does no harm, of course.
3143 6. Moved the call to setjmp() into its own function, to get rid of warnings
3144 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3146 7. Constructs such as \d{8,} were compiling into the equivalent of
3147 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3148 outcome, but in this particular case used more store than had been allocated,
3149 which caused the bug to be discovered because it threw up an internal error.
3151 8. The debugging code in both pcre and pcretest for outputting the compiled
3152 form of a regex was going wrong in the case of back references followed by
3153 curly-bracketed repeats.
3156 Version 1.02 12-Dec-97
3157 ----------------------
3159 1. Typos in pcre.3 and comments in the source fixed.
3161 2. Applied a contributed patch to get rid of places where it used to remove
3162 'const' from variables, and fixed some signed/unsigned and uninitialized
3163 variable warnings.
3165 3. Added the "runtest" target to Makefile.
3167 4. Set default compiler flag to -O2 rather than just -O.
3170 Version 1.01 19-Nov-97
3171 ----------------------
3173 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3174 like /([ab]*)*/, that is, for classes with more than one character in them.
3176 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3177 as /((?>a*))*/ (a PCRE_EXTRA facility).
3180 Version 1.00 18-Nov-97
3181 ----------------------
3183 1. Added compile-time macros to support systems such as SunOS4 which don't have
3184 memmove() or strerror() but have other things that can be used instead.
3186 2. Arranged that "make clean" removes the executables.
3189 Version 0.99 27-Oct-97
3190 ----------------------
3192 1. Fixed bug in code for optimizing classes with only one character. It was
3193 initializing a 32-byte map regardless, which could cause it to run off the end
3194 of the memory it had got.
3196 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3199 Version 0.98 22-Oct-97
3200 ----------------------
3202 1. Fixed bug in code for handling temporary memory usage when there are more
3203 back references than supplied space in the ovector. This could cause segfaults.
3206 Version 0.97 21-Oct-97
3207 ----------------------
3209 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3211 2. Optimized negated single characters not to use a bit map.
3213 3. Brought error texts together as macro definitions; clarified some of them;
3214 fixed one that was wrong - it said "range out of order" when it meant "invalid
3215 escape sequence".
3217 4. Changed some char * arguments to const char *.
3219 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3221 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3222 pcretest.
3225 Version 0.96 16-Oct-97
3226 ----------------------
3228 1. Added a simple "pgrep" utility to the distribution.
3230 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3231 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3232 where "ddd" means "one or more decimal digits".
3234 3. Fixed serious bug. If a pattern had a back reference, but the call to
3235 pcre_exec() didn't supply a large enough ovector to record the related
3236 identifying subpattern, the match always failed. PCRE now remembers the number
3237 of the largest back reference, and gets some temporary memory in which to save
3238 the offsets during matching if necessary, in order to ensure that
3239 backreferences always work.
3241 4. Increased the compatibility with Perl in a number of ways:
3243 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3244 to request this handling. The option can be set at compile or exec time.
3246 (b) $ matches before a terminating newline by default; an option
3247 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3248 mode). The option can be set at compile or exec time.
3250 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3251 the same as Perl's. If the decimal number it represents is less than 10
3252 or there aren't that many previous left capturing parentheses, an octal
3253 escape is read. Inside a character class, it's always an octal escape,
3254 even if it is a single digit.
3256 (d) An escaped but undefined alphabetic character is taken as a literal,
3257 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3258 escapes.
3260 (e) {0} is now permitted. (The previous item is removed from the compiled
3261 pattern).
3263 5. Changed all the names of code files so that the basic parts are no longer
3264 than 10 characters, and abolished the teeny "globals.c" file.
3266 6. Changed the handling of character classes; they are now done with a 32-byte
3267 bit map always.
3269 7. Added the -d and /D options to pcretest to make it possible to look at the
3270 internals of compilation without having to recompile pcre.
3273 Version 0.95 23-Sep-97
3274 ----------------------
3276 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3277 \x20 at the start of a run of normal characters. These were being treated as
3278 real characters, instead of the source characters being re-checked.
3281 Version 0.94 18-Sep-97
3282 ----------------------
3284 1. The functions are now thread-safe, with the caveat that the global variables
3285 containing pointers to malloc() and free() or alternative functions are the
3286 same for all threads.
3288 2. Get pcre_study() to generate a bitmap of initial characters for non-
3289 anchored patterns when this is possible, and use it if passed to pcre_exec().
3292 Version 0.93 15-Sep-97
3293 ----------------------
3295 1. /(b)|(:+)/ was computing an incorrect first character.
3297 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3298 but not actually doing anything yet.
3300 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3301 as Perl does (e.g. [-az] or [az-]).
3303 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3304 all possible positions.
3306 5. Split up into different modules to avoid including unneeded functions in a
3307 compiled binary. However, compile and exec are still in one module. The "study"
3308 function is split off.
3310 6. The character tables are now in a separate module whose source is generated
3311 by an auxiliary program - but can then be edited by hand if required. There are
3312 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3313 toupper() in the code.
3315 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3316 make them global. Abolish the function for setting them, as the caller can now
3317 set them directly.
3320 Version 0.92 11-Sep-97
3321 ----------------------
3323 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3324 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3326 2. Caseless matching was not working in character classes if the characters in
3327 the pattern were in upper case.
3329 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3331 4. Make PCRE_ANCHORED public and accept as a compile option.
3333 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3334 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3335 pass them.
3337 6. Give an error if bad option bits passed at compile or run time.
3339 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3340 pcretest to cause it to pass that flag.
3342 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3343 options, and the first character, if set.
3345 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3348 Version 0.91 10-Sep-97
3349 ----------------------
3351 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3352 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3354 2. PCRE was looping on encountering an indefinitely repeated back reference to
3355 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3356 Perl does - treats the match as successful.
3358 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5