ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 368 - (show annotations)
Sun Aug 24 16:25:20 2008 UTC (13 years, 1 month ago) by ph10
File size: 159458 byte(s)
Error occurred while calculating annotation data.
Make it clearer that ovector values are byte offsets, not character counts.
1 ChangeLog for PCRE
2 ------------------
4 Version 7.8 25-Aug-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.
53 12. For a pattern where the match had to start at the beginning or immediately
54 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
55 pcre_dfa_exec() could read past the end of the passed subject if there was
56 no match. To help with detecting such bugs (e.g. with valgrind), I modified
57 pcretest so that it places the subject at the end of its malloc-ed buffer.
59 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
60 exec() might read past the end of the data buffer in UTF-8 mode.
62 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
63 the data contained the byte 0x85 as part of a UTF-8 character within its
64 first line. This applied both to normal and DFA matching.
66 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
67 /^[^d]*?$/8 failed to match "abc".
69 16. Added a missing copyright notice to pcrecpp_internal.h.
71 17. Make it more clear in the documentation that values returned from
72 pcre_exec() in ovector are byte offsets, not character counts.
75 Version 7.7 07-May-08
76 ---------------------
78 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
79 a string to a long long, pretend we don't even have a long long." This is
80 done by checking for the strtoq, strtoll, and _strtoi64 functions.
82 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
83 pre-7.6 versions, which defined a global no_arg variable instead of putting
84 it in the RE class. (See also #8 below.)
86 3. Remove a line of dead code, identified by coverity and reported by Nuno
87 Lopes.
89 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
91 (1) The include/exclude patterns were being applied to the whole pathnames
92 of files, instead of just to the final components.
94 (2) If there was more than one level of directory, the subdirectories were
95 skipped unless they satisfied the include/exclude conditions. This is
96 inconsistent with GNU grep (and could even be seen as contrary to the
97 pcregrep specification - which I improved to make it absolutely clear).
98 The action now is always to scan all levels of directory, and just
99 apply the include/exclude patterns to regular files.
101 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
102 --exclude_dir in the tests to avoid scanning .svn directories.
104 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
105 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
106 doesn't support NULs in patterns.
108 7. Added some missing "const"s to declarations of static tables in
109 pcre_compile.c and pcre_dfa_exec.c.
111 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
112 caused by fix #2 above. (Subsequently also a second patch to fix the
113 first patch. And a third patch - this was a messy problem.)
115 9. Applied Craig's patch to remove the use of push_back().
117 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
118 matching function regexec().
120 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
121 which, however, unlike Perl's \g{...}, are subroutine calls, not back
122 references. PCRE supports relative numbers with this syntax (I don't think
123 Oniguruma does).
125 12. Previously, a group with a zero repeat such as (...){0} was completely
126 omitted from the compiled regex. However, this means that if the group
127 was called as a subroutine from elsewhere in the pattern, things went wrong
128 (an internal error was given). Such groups are now left in the compiled
129 pattern, with a new opcode that causes them to be skipped at execution
130 time.
132 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
133 to the way PCRE behaves:
135 (a) A lone ] character is dis-allowed (Perl treats it as data).
137 (b) A back reference to an unmatched subpattern matches an empty string
138 (Perl fails the current match path).
140 (c) A data ] in a character class must be notated as \] because if the
141 first data character in a class is ], it defines an empty class. (In
142 Perl it is not possible to have an empty class.) The empty class []
143 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
144 The negative empty class [^] matches any one character, independently
145 of the DOTALL setting.
147 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
148 non-existent subpattern following a character class starting with ']' and
149 containing () gave an internal compiling error instead of "reference to
150 non-existent subpattern". Fortunately, when the pattern did exist, the
151 compiled code was correct. (When scanning forwards to check for the
152 existencd of the subpattern, it was treating the data ']' as terminating
153 the class, so got the count wrong. When actually compiling, the reference
154 was subsequently set up correctly.)
156 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
157 it was being rejected as not supported by pcre_dfa_exec(), even though
158 other assertions are supported. I have made pcre_dfa_exec() support
159 (*FAIL).
161 16. The implementation of 13c above involved the invention of a new opcode,
162 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
163 cannot be changed at match time, I realized I could make a small
164 improvement to matching performance by compiling OP_ALLANY instead of
165 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
166 on the OP_ANY path.
168 17. Compiling pcretest on Windows with readline support failed without the
169 following two fixes: (1) Make the unistd.h include conditional on
170 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
172 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
173 ncurses library to be included for pcretest when ReadLine support is
174 requested, but also to allow for it to be overridden. This patch came from
175 Daniel Bergström.
177 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
178 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
179 any errors with the current Unicode tables. Thanks to Peter Kankowski for
180 spotting this.
183 Version 7.6 28-Jan-08
184 ---------------------
186 1. A character class containing a very large number of characters with
187 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
188 overflow.
190 2. Patch to cut out the "long long" test in pcrecpp_unittest when
191 HAVE_LONG_LONG is not defined.
193 3. Applied Christian Ehrlicher's patch to update the CMake build files to
194 bring them up to date and include new features. This patch includes:
196 - Fixed PH's badly added libz and libbz2 support.
197 - Fixed a problem with static linking.
198 - Added pcredemo. [But later removed - see 7 below.]
199 - Fixed dftables problem and added an option.
200 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
202 - Added readline support for pcretest.
203 - Added an listing of the option settings after cmake has run.
205 4. A user submitted a patch to Makefile that makes it easy to create
206 "pcre.dll" under mingw when using Configure/Make. I added stuff to
207 Makefile.am that cause it to include this special target, without
208 affecting anything else. Note that the same mingw target plus all
209 the other distribution libraries and programs are now supported
210 when configuring with CMake (see 6 below) instead of with
211 Configure/Make.
213 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
214 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
215 exported in the Windows port". It has not yet been confirmed that the patch
216 solves the problem, but it does no harm.
218 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
219 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
220 with CMake, and also correct the comment about stack recursion.
222 7. Remove the automatic building of pcredemo from the ./configure system and
223 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
224 of a program that users should build themselves after PCRE is installed, so
225 building it automatically is not really right. What is more, it gave
226 trouble in some build environments.
228 8. Further tidies to CMakeLists.txt from Sheri and Christian.
231 Version 7.5 10-Jan-08
232 ---------------------
234 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
235 values in parens when parsing an RE using the C++ wrapper."
237 2. Negative specials like \S did not work in character classes in UTF-8 mode.
238 Characters greater than 255 were excluded from the class instead of being
239 included.
241 3. The same bug as (2) above applied to negated POSIX classes such as
242 [:^space:].
244 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
245 defined or documented. It seems to have been a typo for PCRE_STATIC, so
246 I have changed it.
248 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
249 first named subpattern) and a construct such as (?&a) would reference the
250 first named subpattern whose name started with "a" (in other words, the
251 length check was missing). Both these problems are fixed. "Subpattern name
252 expected" is now given for (?&) (a zero-length name), and this patch also
253 makes it give the same error for \k'' (previously it complained that that
254 was a reference to a non-existent subpattern).
256 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
257 this is right because (?- can be followed by option settings as well as by
258 digits. I have, however, made the messages clearer.
260 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
261 than the number used in the conditional) now cause a compile-time error.
262 This is actually not compatible with Perl, which accepts such patterns, but
263 treats the conditional as always being FALSE (as PCRE used to), but it
264 seems to me that giving a diagnostic is better.
266 8. Change "alphameric" to the more common word "alphanumeric" in comments
267 and messages.
269 9. Fix two occurrences of "backslash" in comments that should have been
270 "backspace".
272 10. Remove two redundant lines of code that can never be obeyed (their function
273 was moved elsewhere).
275 11. The program that makes PCRE's Unicode character property table had a bug
276 which caused it to generate incorrect table entries for sequences of
277 characters that have the same character type, but are in different scripts.
278 It amalgamated them into a single range, with the script of the first of
279 them. In other words, some characters were in the wrong script. There were
280 thirteen such cases, affecting characters in the following ranges:
282 U+002b0 - U+002c1
283 U+0060c - U+0060d
284 U+0061e - U+00612
285 U+0064b - U+0065e
286 U+0074d - U+0076d
287 U+01800 - U+01805
288 U+01d00 - U+01d77
289 U+01d9b - U+01dbf
290 U+0200b - U+0200f
291 U+030fc - U+030fe
292 U+03260 - U+0327f
293 U+0fb46 - U+0fbb1
294 U+10450 - U+1049d
296 12. The -o option (show only the matching part of a line) for pcregrep was not
297 compatible with GNU grep in that, if there was more than one match in a
298 line, it showed only the first of them. It now behaves in the same way as
299 GNU grep.
301 13. If the -o and -v options were combined for pcregrep, it printed a blank
302 line for every non-matching line. GNU grep prints nothing, and pcregrep now
303 does the same. The return code can be used to tell if there were any
304 non-matching lines.
306 14. Added --file-offsets and --line-offsets to pcregrep.
308 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
309 infinitely looping recursion. The bug was that positive lookaheads were not
310 being skipped when checking for a possible empty match (negative lookaheads
311 and both kinds of lookbehind were skipped).
313 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
314 inclusion of <windows.h> to before rather than after the definition of
315 INVALID_FILE_ATTRIBUTES (patch from David Byron).
317 17. Specifying a possessive quantifier with a specific limit for a Unicode
318 character property caused pcre_compile() to compile bad code, which led at
319 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
320 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
321 caused the error; without that there was no problem.
323 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
325 19. Added --enable-pcretest-libreadline.
327 20. In pcrecpp.cc, the variable 'count' was incremented twice in
328 RE::GlobalReplace(). As a result, the number of replacements returned was
329 double what it should be. I removed one of the increments, but Craig sent a
330 later patch that removed the other one (the right fix) and added unit tests
331 that check the return values (which was not done before).
333 21. Several CMake things:
335 (1) Arranged that, when cmake is used on Unix, the libraries end up with
336 the names libpcre and libpcreposix, not just pcre and pcreposix.
338 (2) The above change means that pcretest and pcregrep are now correctly
339 linked with the newly-built libraries, not previously installed ones.
343 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
344 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
345 UTF-8 newline character). The key issue is that the pattern starts .*;
346 this means that the match must be either at the beginning, or after a
347 newline. The bug was in the code for advancing after a failed match and
348 checking that the new position followed a newline. It was not taking
349 account of UTF-8 characters correctly.
351 23. PCRE was behaving differently from Perl in the way it recognized POSIX
352 character classes. PCRE was not treating the sequence [:...:] as a
353 character class unless the ... were all letters. Perl, however, seems to
354 allow any characters between [: and :], though of course it rejects as
355 unknown any "names" that contain non-letters, because all the known class
356 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
357 for example, whereas PCRE did not - it did not recognize a POSIX character
358 class. This seemed a bit dangerous, so the code has been changed to be
359 closer to Perl. The behaviour is not identical to Perl, because PCRE will
360 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
361 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
362 Perl does, and where it didn't before.
364 24. Rewrite so as to remove the single use of %n from pcregrep because in some
365 Windows environments %n is disabled by default.
368 Version 7.4 21-Sep-07
369 ---------------------
371 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
372 means that a class such as [\s] counted as "explicit reference to CR or
373 LF". That isn't really right - the whole point of the change was to try to
374 help when there was an actual mention of one of the two characters. So now
375 the change happens only if \r or \n (or a literal CR or LF) character is
376 encountered.
378 2. The 32-bit options word was also used for 6 internal flags, but the numbers
379 of both had grown to the point where there were only 3 bits left.
380 Fortunately, there was spare space in the data structure, and so I have
381 moved the internal flags into a new 16-bit field to free up more option
382 bits.
384 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
385 but did not set the internal JCHANGED flag - either of these is enough to
386 control the way the "get" function works - but the PCRE_INFO_JCHANGED
387 facility is supposed to tell if (?J) was ever used, so now (?J) at the
388 start sets both bits.
390 4. Added options (at build time, compile time, exec time) to change \R from
391 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
393 5. doc/pcresyntax.html was missing from the distribution.
395 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
396 compatibility, even though it is no longer used.
398 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
399 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
400 windows.h file is present (where different names are used). [This was
401 reversed later after testing - see 16 below.]
403 8. Changed all #include <config.h> to #include "config.h". There were also
404 some further <pcre.h> cases that I changed to "pcre.h".
406 9. When pcregrep was used with the --colour option, it missed the line ending
407 sequence off the lines that it output.
409 10. It was pointed out to me that arrays of string pointers cause lots of
410 relocations when a shared library is dynamically loaded. A technique of
411 using a single long string with a table of offsets can drastically reduce
412 these. I have refactored PCRE in four places to do this. The result is
413 dramatic:
415 Originally: 290
416 After changing UCP table: 187
417 After changing error message table: 43
418 After changing table of "verbs" 36
419 After changing table of Posix names 22
421 Thanks to the folks working on Gregex for glib for this insight.
423 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
424 unicode-properties was also set.
426 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
428 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
429 checked only for CRLF.
431 14. Added casts to pcretest.c to avoid compiler warnings.
433 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
435 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
436 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
437 entirely. This removes changes made in 7 above.
439 17. The CMake files have been updated, and there is now more information about
440 building with CMake in the NON-UNIX-USE document.
443 Version 7.3 28-Aug-07
444 ---------------------
446 1. In the rejigging of the build system that eventually resulted in 7.1, the
447 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
448 brackets there is not right, since it causes compilers to look for an
449 installed pcre.h, not the version that is in the source that is being
450 compiled (which of course may be different). I have changed it back to:
452 #include "pcre.h"
454 I have a vague recollection that the change was concerned with compiling in
455 different directories, but in the new build system, that is taken care of
456 by the VPATH setting the Makefile.
458 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
459 when the subject happened to end in the byte 0x85 (e.g. if the last
460 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
461 characters but of course it shouldn't be taken as a newline when it is part
462 of another character. The bug was that, for an unlimited repeat of . in
463 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
464 characters when looking for a newline.
466 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
468 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
469 in debug output.
471 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
472 long printing in the pcrecpp unittest when running under MinGW.
474 6. ESC_K was left out of the EBCDIC table.
476 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
477 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
478 limit also applies to "virtual nesting" when a pattern is recursive, and in
479 this case 1000 isn't so big. I have been able to remove this limit at the
480 expense of backing off one optimization in certain circumstances. Normally,
481 when pcre_exec() would call its internal match() function recursively and
482 immediately return the result unconditionally, it uses a "tail recursion"
483 feature to save stack. However, when a subpattern that can match an empty
484 string has an unlimited repetition quantifier, it no longer makes this
485 optimization. That gives it a stack frame in which to save the data for
486 checking that an empty string has been matched. Previously this was taken
487 from the 1000-entry workspace that had been reserved. So now there is no
488 explicit limit, but more stack is used.
490 8. Applied Daniel's patches to solve problems with the import/export magic
491 syntax that is required for Windows, and which was going wrong for the
492 pcreposix and pcrecpp parts of the library. These were overlooked when this
493 problem was solved for the main library.
495 9. There were some crude static tests to avoid integer overflow when computing
496 the size of patterns that contain repeated groups with explicit upper
497 limits. As the maximum quantifier is 65535, the maximum group length was
498 set at 30,000 so that the product of these two numbers did not overflow a
499 32-bit integer. However, it turns out that people want to use groups that
500 are longer than 30,000 bytes (though not repeat them that many times).
501 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
502 made it possible to implement the integer overflow checks in a much more
503 dynamic way, which I have now done. The artificial limitation on group
504 length has been removed - we now have only the limit on the total length of
505 the compiled pattern, which depends on the LINK_SIZE setting.
507 10. Fixed a bug in the documentation for get/copy named substring when
508 duplicate names are permitted. If none of the named substrings are set, the
509 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
510 empty string.
512 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
513 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
514 because the ] is interpreted as the first data character and the
515 terminating ] is not found. PCRE has been made compatible with Perl in this
516 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
517 cause memory overwriting.
519 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
520 string has been matched (to stop an infinite loop). It was not recognizing
521 a conditional subpattern that could match an empty string if that
522 subpattern was within another subpattern. For example, it looped when
523 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
524 condition was not nested. This bug has been fixed.
526 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
527 past the start of the subject in the presence of bytes with the top bit
528 set, for example "\x8aBCD".
530 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
531 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
533 14. Optimized (?!) to (*FAIL).
535 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
536 This restricts code points to be within the range 0 to 0x10FFFF, excluding
537 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
538 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
539 does: it's just the validity check that is more restrictive.
541 16. Inserted checks for integer overflows during escape sequence (backslash)
542 processing, and also fixed erroneous offset values for syntax errors during
543 backslash processing.
545 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
546 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
548 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
549 caused an overrun.
551 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
552 something other than just ASCII characters) inside a group that had an
553 unlimited repeat caused a loop at compile time (while checking to see
554 whether the group could match an empty string).
556 20. Debugging a pattern containing \p or \P could cause a crash. For example,
557 [\P{Any}] did so. (Error in the code for printing property names.)
559 21. An orphan \E inside a character class could cause a crash.
561 22. A repeated capturing bracket such as (A)? could cause a wild memory
562 reference during compilation.
564 23. There are several functions in pcre_compile() that scan along a compiled
565 expression for various reasons (e.g. to see if it's fixed length for look
566 behind). There were bugs in these functions when a repeated \p or \P was
567 present in the pattern. These operators have additional parameters compared
568 with \d, etc, and these were not being taken into account when moving along
569 the compiled data. Specifically:
571 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
572 length.
574 (b) An item such as \pL+ within a repeated group could cause crashes or
575 loops.
577 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
578 "reference to non-existent subpattern" error.
580 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
582 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
583 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
585 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
587 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
588 character were causing crashes (broken optimization).
590 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
591 \p or \P) caused a compile-time loop.
593 28. More problems have arisen in unanchored patterns when CRLF is a valid line
594 break. For example, the unstudied pattern [\r\n]A does not match the string
595 "\r\nA" because change 7.0/46 below moves the current point on by two
596 characters after failing to match at the start. However, the pattern \nA
597 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
598 the same is true. There doesn't seem any very clean way out of this, but
599 what I have chosen to do makes the common cases work: PCRE now takes note
600 of whether there can be an explicit match for \r or \n anywhere in the
601 pattern, and if so, 7.0/46 no longer applies. As part of this change,
602 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
603 pattern has explicit CR or LF references.
605 29. Added (*CR) etc for changing newline setting at start of pattern.
608 Version 7.2 19-Jun-07
609 ---------------------
611 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
612 which is apparently normally available under Windows.
614 2. Re-jig the pcregrep tests with different newline settings in an attempt
615 to make them independent of the local environment's newline setting.
617 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
619 4. Some of the "internals" tests were previously cut out when the link size
620 was not 2, because the output contained actual offsets. The recent new
621 "Z" feature of pcretest means that these can be cut out, making the tests
622 usable with all link sizes.
624 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
625 stack recursion. This gives a massive performance boost under BSD, but just
626 a small improvement under Linux. However, it saves one field in the frame
627 in all cases.
629 6. Added more features from the forthcoming Perl 5.10:
631 (a) (?-n) (where n is a string of digits) is a relative subroutine or
632 recursion call. It refers to the nth most recently opened parentheses.
634 (b) (?+n) is also a relative subroutine call; it refers to the nth next
635 to be opened parentheses.
637 (c) Conditions that refer to capturing parentheses can be specified
638 relatively, for example, (?(-2)... or (?(+3)...
640 (d) \K resets the start of the current match so that everything before
641 is not part of it.
643 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
645 (f) \g{name} is another synonym - part of Perl 5.10's unification of
646 reference syntax.
648 (g) (?| introduces a group in which the numbering of parentheses in each
649 alternative starts with the same number.
651 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
653 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
656 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
657 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
658 for detecting groups that can match an empty string.
660 9. A pattern with a very large number of alternatives (more than several
661 hundred) was running out of internal workspace during the pre-compile
662 phase, where pcre_compile() figures out how much memory will be needed. A
663 bit of new cunning has reduced the workspace needed for groups with
664 alternatives. The 1000-alternative test pattern now uses 12 bytes of
665 workspace instead of running out of the 4096 that are available.
667 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
669 11. Applied patch from Google to remove an optimization that didn't quite work.
670 The report of the bug said:
672 pcrecpp::RE("a*").FullMatch("aaa") matches, while
673 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
674 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
676 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
677 it matched the wrong number of bytes.
680 Version 7.1 24-Apr-07
681 ---------------------
683 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
684 that is more "standard", making use of automake and other Autotools. There
685 is some re-arrangement of the files and adjustment of comments consequent
686 on this.
688 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
689 for recursive directory scanning broke on some systems because the files
690 are not scanned in any specific order and on different systems the order
691 was different. A call to "sort" has been inserted into RunGrepTest for the
692 approprate test as a short-term fix. In the longer term there may be an
693 alternative.
695 3. I had an email from Eric Raymond about problems translating some of PCRE's
696 man pages to HTML (despite the fact that I distribute HTML pages, some
697 people do their own conversions for various reasons). The problems
698 concerned the use of low-level troff macros .br and .in. I have therefore
699 removed all such uses from the man pages (some were redundant, some could
700 be replaced by .nf/.fi pairs). The 132html script that I use to generate
701 HTML has been updated to handle .nf/.fi and to complain if it encounters
702 .br or .in.
704 4. Updated comments in configure.ac that get placed in config.h.in and also
705 arranged for config.h to be included in the distribution, with the name
706 config.h.generic, for the benefit of those who have to compile without
707 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
709 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
710 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
711 makevp.bat for the current PCRE, using the additional files
712 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
714 6. A Windows user reported a minor discrepancy with test 2, which turned out
715 to be caused by a trailing space on an input line that had got lost in his
716 copy. The trailing space was an accident, so I've just removed it.
718 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
719 that is needed.
721 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
722 as "const" (a) because they are and (b) because it helps the PHP
723 maintainers who have recently made a script to detect big data structures
724 in the php code that should be moved to the .rodata section. I remembered
725 to update Builducptable as well, so it won't revert if ucptable.h is ever
726 re-created.
728 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
729 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
730 order to be able to cut out the UTF-8 tables in the latter when UTF-8
731 support is not required. This saves 1.5-2K of code, which is important in
732 some applications.
734 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
735 so as not to refer to the tables, even though these functions will never be
736 called when UTF-8 support is disabled. Otherwise there are problems with a
737 shared library.
739 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
741 (a) It was defining its arguments as char * instead of void *.
743 (b) It was assuming that all moves were upwards in memory; this was true
744 a long time ago when I wrote it, but is no longer the case.
746 The emulated memove() is provided for those environments that have neither
747 memmove() nor bcopy(). I didn't think anyone used it these days, but that
748 is clearly not the case, as these two bugs were recently reported.
750 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
751 and Detrail to create the HTML documentation, the .txt form of the man
752 pages, and it removes trailing spaces from listed files. It also creates
753 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
754 case, it wraps all the #defines with #ifndefs. This script should be run
755 before "make dist".
757 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
758 with Unicode property support.
760 (a) For a maximizing quantifier, if the two different cases of the
761 character were of different lengths in their UTF-8 codings (there are
762 some cases like this - I found 11), and the matching function had to
763 back up over a mixture of the two cases, it incorrectly assumed they
764 were both the same length.
766 (b) When PCRE was configured to use the heap rather than the stack for
767 recursion during matching, it was not correctly preserving the data for
768 the other case of a UTF-8 character when checking ahead for a match
769 while processing a minimizing repeat. If the check also involved
770 matching a wide character, but failed, corruption could cause an
771 erroneous result when trying to check for a repeat of the original
772 character.
774 13. Some tidying changes to the testing mechanism:
776 (a) The RunTest script now detects the internal link size and whether there
777 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
778 values substituted by "configure". (The RunGrepTest script already did
779 this for UTF-8.) The configure.ac script no longer substitutes the
780 relevant variables.
782 (b) The debugging options /B and /D in pcretest show the compiled bytecode
783 with length and offset values. This means that the output is different
784 for different internal link sizes. Test 2 is skipped for link sizes
785 other than 2 because of this, bypassing the problem. Unfortunately,
786 there was also a test in test 3 (the locale tests) that used /B and
787 failed for link sizes other than 2. Rather than cut the whole test out,
788 I have added a new /Z option to pcretest that replaces the length and
789 offset values with spaces. This is now used to make test 3 independent
790 of link size. (Test 2 will be tidied up later.)
792 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
793 segmentation fault instead of returning the appropriate error message.
795 15. In multiline mode when the newline sequence was set to "any", the pattern
796 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
797 This doesn't seem right; it now treats the CRLF combination as the line
798 ending, and so does not match in that case. It's only a pattern such as ^$
799 that would hit this one: something like ^ABC$ would have failed after \r
800 and then tried again after \r\n.
802 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
803 in an attempt to make files that differ only in their line terminators
804 compare equal. This works on Linux.
806 17. Under certain error circumstances pcregrep might try to free random memory
807 as it exited. This is now fixed, thanks to valgrind.
809 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
810 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
811 was because its rules for how to advance for /g after matching an empty
812 string at the end of a line did not allow for this case. They now check for
813 it specially.
815 20. pcretest is supposed to handle patterns and data of any length, by
816 extending its buffers when necessary. It was getting this wrong when the
817 buffer for a data line had to be extended.
819 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
820 CRLF as a newline sequence.
822 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
823 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
824 I have nevertheless tidied it up.
826 23. Added some casts to kill warnings from HP-UX ia64 compiler.
828 24. Added a man page for pcre-config.
831 Version 7.0 19-Dec-06
832 ---------------------
834 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
835 moving to gcc 4.1.1.
837 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
838 sys/time.h, which is documented as needed for this function. It doesn't
839 seem to matter on Linux, but it showed up on some releases of OS X.
841 3. It seems that there are systems where bytes whose values are greater than
842 127 match isprint() in the "C" locale. The "C" locale should be the
843 default when a C program starts up. In most systems, only ASCII printing
844 characters match isprint(). This difference caused the output from pcretest
845 to vary, making some of the tests fail. I have changed pcretest so that:
847 (a) When it is outputting text in the compiled version of a pattern, bytes
848 other than 32-126 are always shown as hex escapes.
850 (b) When it is outputting text that is a matched part of a subject string,
851 it does the same, unless a different locale has been set for the match
852 (using the /L modifier). In this case, it uses isprint() to decide.
854 4. Fixed a major bug that caused incorrect computation of the amount of memory
855 required for a compiled pattern when options that changed within the
856 pattern affected the logic of the preliminary scan that determines the
857 length. The relevant options are -x, and -i in UTF-8 mode. The result was
858 that the computed length was too small. The symptoms of this bug were
859 either the PCRE error "internal error: code overflow" from pcre_compile(),
860 or a glibc crash with a message such as "pcretest: free(): invalid next
861 size (fast)". Examples of patterns that provoked this bug (shown in
862 pcretest format) are:
864 /(?-x: )/x
865 /(?x)(?-x: \s*#\s*)/
866 /((?i)[\x{c0}])/8
867 /(?i:[\x{c0}])/8
869 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
870 is now done differently.
872 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
873 wrapper classes; (b) implement a new function in the C++ scanner that is
874 more efficient than the old way of doing things because it avoids levels of
875 recursion in the regex matching; (c) add a paragraph to the documentation
876 for the FullMatch() function.
878 6. The escape sequence \n was being treated as whatever was defined as
879 "newline". Not only was this contrary to the documentation, which states
880 that \n is character 10 (hex 0A), but it also went horribly wrong when
881 "newline" was defined as CRLF. This has been fixed.
883 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
884 was being set to -1 for the "end of line" case (supposedly a value that no
885 character can have). Though this value is never used (the check for end of
886 line is "zero bytes in current character"), it caused compiler complaints.
887 I've changed it to 0xffffffff.
889 8. In pcre_version.c, the version string was being built by a sequence of
890 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
891 string (as it is for production releases) called a macro with an empty
892 argument. The C standard says the result of this is undefined. The gcc
893 compiler treats it as an empty string (which was what was wanted) but it is
894 reported that Visual C gives an error. The source has been hacked around to
895 avoid this problem.
897 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
898 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
899 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
900 of them did).
902 10. Originally, pcretest opened its input and output without "b"; then I was
903 told that "b" was needed in some environments, so it was added for release
904 5.0 to both the input and output. (It makes no difference on Unix-like
905 systems.) Later I was told that it is wrong for the input on Windows. I've
906 now abstracted the modes into two macros, to make it easier to fiddle with
907 them, and removed "b" from the input mode under Windows.
909 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
911 12. Added -help and --help to pcretest as an official way of being reminded
912 of the options.
914 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
915 and pcrecpp.cc because they annoy compilers at high warning levels.
917 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
919 15. Fixed an occurrence of == in configure.ac that should have been = (shell
920 scripts are not C programs :-) and which was not noticed because it works
921 on Linux.
923 16. pcretest is supposed to handle any length of pattern and data line (as one
924 line or as a continued sequence of lines) by extending its input buffer if
925 necessary. This feature was broken for very long pattern lines, leading to
926 a string of junk being passed to pcre_compile() if the pattern was longer
927 than about 50K.
929 17. I have done a major re-factoring of the way pcre_compile() computes the
930 amount of memory needed for a compiled pattern. Previously, there was code
931 that made a preliminary scan of the pattern in order to do this. That was
932 OK when PCRE was new, but as the facilities have expanded, it has become
933 harder and harder to keep it in step with the real compile phase, and there
934 have been a number of bugs (see for example, 4 above). I have now found a
935 cunning way of running the real compile function in a "fake" mode that
936 enables it to compute how much memory it would need, while actually only
937 ever using a few hundred bytes of working memory and without too many
938 tests of the mode. This should make future maintenance and development
939 easier. A side effect of this work is that the limit of 200 on the nesting
940 depth of parentheses has been removed (though this was never a serious
941 limitation, I suspect). However, there is a downside: pcre_compile() now
942 runs more slowly than before (30% or more, depending on the pattern). I
943 hope this isn't a big issue. There is no effect on runtime performance.
945 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
946 newline (only possible for the last line of a file) and it was a
947 pattern that set a locale (followed by /Lsomething), pcretest crashed.
949 19. Added additional timing features to pcretest. (1) The -tm option now times
950 matching only, not compiling. (2) Both -t and -tm can be followed, as a
951 separate command line item, by a number that specifies the number of
952 repeats to use when timing. The default is 50000; this gives better
953 precision, but takes uncomfortably long for very large patterns.
955 20. Extended pcre_study() to be more clever in cases where a branch of a
956 subpattern has no definite first character. For example, (a*|b*)[cd] would
957 previously give no result from pcre_study(). Now it recognizes that the
958 first character must be a, b, c, or d.
960 21. There was an incorrect error "recursive call could loop indefinitely" if
961 a subpattern (or the entire pattern) that was being tested for matching an
962 empty string contained only one non-empty item after a nested subpattern.
963 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
964 incorrectly, because the \d was being skipped in the check.
966 22. The pcretest program now has a new pattern option /B and a command line
967 option -b, which is equivalent to adding /B to every pattern. This causes
968 it to show the compiled bytecode, without the additional information that
969 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
970 is the same as /B/I).
972 23. A new optimization is now able automatically to treat some sequences such
973 as a*b as a*+b. More specifically, if something simple (such as a character
974 or a simple class like \d) has an unlimited quantifier, and is followed by
975 something that cannot possibly match the quantified thing, the quantifier
976 is automatically "possessified".
978 24. A recursive reference to a subpattern whose number was greater than 39
979 went wrong under certain circumstances in UTF-8 mode. This bug could also
980 have affected the operation of pcre_study().
982 25. Realized that a little bit of performance could be had by replacing
983 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
985 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
987 27. Possessive quantifiers such as a++ were previously implemented by turning
988 them into atomic groups such as ($>a+). Now they have their own opcodes,
989 which improves performance. This includes the automatically created ones
990 from 23 above.
992 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
993 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
994 the first matched character to be a colon. This applied both to named and
995 numbered groups.
997 29. The ucpinternal.h header file was missing its idempotency #ifdef.
999 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1000 building PCRE on Windows easier, so I have included it in the distribution.
1002 31. There is now a check in pcretest against a ridiculously large number being
1003 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1004 loop, the loop is abandoned.
1006 32. Forward references to subpatterns in conditions such as (?(2)...) where
1007 subpattern 2 is defined later cause pcre_compile() to search forwards in
1008 the pattern for the relevant set of parentheses. This search went wrong
1009 when there were unescaped parentheses in a character class, parentheses
1010 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1012 33. "Subroutine" calls and backreferences were previously restricted to
1013 referencing subpatterns earlier in the regex. This restriction has now
1014 been removed.
1016 34. Added a number of extra features that are going to be in Perl 5.10. On the
1017 whole, these are just syntactic alternatives for features that PCRE had
1018 previously implemented using the Python syntax or my own invention. The
1019 other formats are all retained for compatibility.
1021 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1022 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1023 also .NET compatible.
1025 (b) A recursion or subroutine call to a named group can now be defined as
1026 (?&name) as well as (?P>name).
1028 (c) A backreference to a named group can now be defined as \k<name> or
1029 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1030 5.10, are also .NET compatible.
1032 (d) A conditional reference to a named group can now use the syntax
1033 (?(<name>) or (?('name') as well as (?(name).
1035 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1036 groups (named and numbered) that are never evaluated inline, but can be
1037 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1038 is always false. There may be only one alternative in such a group.
1040 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1041 as the simple (?(R). The condition is true only if the most recent
1042 recursion is that of the given number or name. It does not search out
1043 through the entire recursion stack.
1045 (g) The escape \gN or \g{N} has been added, where N is a positive or
1046 negative number, specifying an absolute or relative reference.
1048 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1049 some "unreachable code" warnings.
1051 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1052 things, this adds five new scripts.
1054 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1055 There were also incompatibilities regarding the handling of \Q..\E inside
1056 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1057 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1059 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1060 matches an empty string, and forcibly breaks the loop. There were bugs in
1061 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1062 against aaaa the result was just "a" rather than "aaaa", for example. Two
1063 separate and independent bugs (that affected different cases) have been
1064 fixed.
1066 39. Refactored the code to abolish the use of different opcodes for small
1067 capturing bracket numbers. This is a tidy that I avoided doing when I
1068 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1069 The new approach is not only tidier, it makes it possible to reduce the
1070 memory needed to fix the previous bug (38).
1072 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1073 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1074 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1075 mode.
1077 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1078 report.
1080 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1081 copy construction and assignment in the C++ wrapper.
1083 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1084 couple of bugs that could have given wrong results in the "--newline=crlf"
1085 case.
1087 44. Added a number of casts and did some reorganization of signed/unsigned int
1088 variables following suggestions from Dair Grant. Also renamed the variable
1089 "this" as "item" because it is a C++ keyword.
1091 45. Arranged for dftables to add
1093 #include "pcre_internal.h"
1095 to pcre_chartables.c because without it, gcc 4.x may remove the array
1096 definition from the final binary if PCRE is built into a static library and
1097 dead code stripping is activated.
1099 46. For an unanchored pattern, if a match attempt fails at the start of a
1100 newline sequence, and the newline setting is CRLF or ANY, and the next two
1101 characters are CRLF, advance by two characters instead of one.
1104 Version 6.7 04-Jul-06
1105 ---------------------
1107 1. In order to handle tests when input lines are enormously long, pcretest has
1108 been re-factored so that it automatically extends its buffers when
1109 necessary. The code is crude, but this _is_ just a test program. The
1110 default size has been increased from 32K to 50K.
1112 2. The code in pcre_study() was using the value of the re argument before
1113 testing it for NULL. (Of course, in any sensible call of the function, it
1114 won't be NULL.)
1116 3. The memmove() emulation function in pcre_internal.h, which is used on
1117 systems that lack both memmove() and bcopy() - that is, hardly ever -
1118 was missing a "static" storage class specifier.
1120 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1121 containing an extended class (one that cannot be represented by a bitmap
1122 because it contains high-valued characters or Unicode property items, e.g.
1123 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1124 pattern, but PCRE should not loop if you do not (it no longer does).
1125 [Detail: two cases were found: (a) a repeated subpattern containing an
1126 extended class; (b) a recursive reference to a subpattern that followed a
1127 previous extended class. It wasn't skipping over the extended class
1128 correctly when UTF-8 mode was not set.]
1130 5. A negated single-character class was not being recognized as fixed-length
1131 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1132 compile error "lookbehind assertion is not fixed length".
1134 6. The RunPerlTest auxiliary script was showing an unexpected difference
1135 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1136 write a Perl script that can interpret lines of an input file either as
1137 byte characters or as UTF-8, which is what "perltest" was being required to
1138 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1139 can't do is switch easily at run time between having the "use utf8;" pragma
1140 or not. In the end, I fudged it by using the RunPerlTest script to insert
1141 "use utf8;" explicitly for the UTF-8 tests.
1143 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1144 the end of the subject string, contrary to the documentation and to what
1145 Perl does. This was true of both matching functions. Now it matches only at
1146 the start of the subject and immediately after *internal* newlines.
1148 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1149 a pointer to an int instead of a pointer to an unsigned long int. This
1150 caused problems on 64-bit systems.
1152 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1153 instance of the 'standard' template library not being so standard".
1155 10. There was no check on the number of named subpatterns nor the maximum
1156 length of a subpattern name. The product of these values is used to compute
1157 the size of the memory block for a compiled pattern. By supplying a very
1158 long subpattern name and a large number of named subpatterns, the size
1159 computation could be caused to overflow. This is now prevented by limiting
1160 the length of names to 32 characters, and the number of named subpatterns
1161 to 10,000.
1163 11. Subpatterns that are repeated with specific counts have to be replicated in
1164 the compiled pattern. The size of memory for this was computed from the
1165 length of the subpattern and the repeat count. The latter is limited to
1166 65535, but there was no limit on the former, meaning that integer overflow
1167 could in principle occur. The compiled length of a repeated subpattern is
1168 now limited to 30,000 bytes in order to prevent this.
1170 12. Added the optional facility to have named substrings with the same name.
1172 13. Added the ability to use a named substring as a condition, using the
1173 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1174 are numbers (not recommended). Forward references are permitted.
1176 14. Added forward references in named backreferences (if you see what I mean).
1178 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1179 pattern could run off the end of the subject. For example, the pattern
1180 "(?s)(.{1,5})"8 did this with the subject "ab".
1182 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1183 PCRE_CASELESS was set when matching characters that were quantified with ?
1184 or *.
1186 17. A character class other than a single negated character that had a minimum
1187 but no maximum quantifier - for example [ab]{6,} - was not handled
1188 correctly by pce_dfa_exec(). It would match only one character.
1190 18. A valid (though odd) pattern that looked like a POSIX character
1191 class but used an invalid character after [ (for example [[,abc,]]) caused
1192 pcre_compile() to give the error "Failed: internal error: code overflow" or
1193 in some cases to crash with a glibc free() error. This could even happen if
1194 the pattern terminated after [[ but there just happened to be a sequence of
1195 letters, a binary zero, and a closing ] in the memory that followed.
1197 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1198 over the years. Originally (before any Unicode support), just the bottom 8
1199 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1200 output from "man perlunicode" includes this:
1202 The regular expression compiler produces polymorphic opcodes. That
1203 is, the pattern adapts to the data and automatically switches to
1204 the Unicode character scheme when presented with Unicode data--or
1205 instead uses a traditional byte scheme when presented with byte
1206 data.
1208 Sadly, a wide octal escape does not cause a switch, and in a string with
1209 no other multibyte characters, these octal escapes are treated as before.
1210 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1211 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1212 Unicode string.
1214 I have not perpetrated such confusion in PCRE. Up till now, it took just
1215 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1216 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1217 translate to the appropriate multibyte character.
1219 29. Applied some refactoring to reduce the number of warnings from Microsoft
1220 and Borland compilers. This has included removing the fudge introduced
1221 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1222 a warning about an unused variable.
1224 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1225 characters since release 4.0, because Perl (from release 5.004) does not.
1226 [Or at least, is documented not to: some releases seem to be in conflict
1227 with the documentation.] However, when a pattern was studied with
1228 pcre_study() and all its branches started with \s, PCRE still included VT
1229 as a possible starting character. Of course, this did no harm; it just
1230 caused an unnecessary match attempt.
1232 22. Removed a now-redundant internal flag bit that recorded the fact that case
1233 dependency changed within the pattern. This was once needed for "required
1234 byte" processing, but is no longer used. This recovers a now-scarce options
1235 bit. Also moved the least significant internal flag bit to the most-
1236 significant bit of the word, which was not previously used (hangover from
1237 the days when it was an int rather than a uint) to free up another bit for
1238 the future.
1240 23. Added support for CRLF line endings as well as CR and LF. As well as the
1241 default being selectable at build time, it can now be changed at runtime
1242 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1243 specify that it is scanning data with non-default line endings.
1245 24. Changed the definition of CXXLINK to make it agree with the definition of
1246 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1248 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1249 recursions. This makes a big different to stack usage for some patterns.
1251 26. If a subpattern containing a named recursion or subroutine reference such
1252 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1253 the space required for the compiled pattern went wrong and gave too small a
1254 value. Depending on the environment, this could lead to "Failed: internal
1255 error: code overflow at offset 49" or "glibc detected double free or
1256 corruption" errors.
1258 27. Applied patches from Google (a) to support the new newline modes and (b) to
1259 advance over multibyte UTF-8 characters in GlobalReplace.
1261 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1262 difference for some implementation of PCRE in some Windows version.
1264 29. Added some extra testing facilities to pcretest:
1266 \q<number> in a data line sets the "match limit" value
1267 \Q<number> in a data line sets the "match recursion limt" value
1268 -S <number> sets the stack size, where <number> is in megabytes
1270 The -S option isn't available for Windows.
1273 Version 6.6 06-Feb-06
1274 ---------------------
1276 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1277 in pcreposix.h. I have copied the definition from pcre.h.
1279 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1280 because pcre.h is no longer a built file.
1282 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1283 not normally included in the compiled code.
1286 Version 6.5 01-Feb-06
1287 ---------------------
1289 1. When using the partial match feature with pcre_dfa_exec(), it was not
1290 anchoring the second and subsequent partial matches at the new starting
1291 point. This could lead to incorrect results. For example, with the pattern
1292 /1234/, partially matching against "123" and then "a4" gave a match.
1294 2. Changes to pcregrep:
1296 (a) All non-match returns from pcre_exec() were being treated as failures
1297 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1298 error message is output. Some extra information is given for the
1300 probably the only errors that are likely to be caused by users (by
1301 specifying a regex that has nested indefinite repeats, for instance).
1302 If there are more than 20 of these errors, pcregrep is abandoned.
1304 (b) A binary zero was treated as data while matching, but terminated the
1305 output line if it was written out. This has been fixed: binary zeroes
1306 are now no different to any other data bytes.
1308 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1309 used to set a locale for matching. The --locale=xxxx long option has
1310 been added (no short equivalent) to specify a locale explicitly on the
1311 pcregrep command, overriding the environment variables.
1313 (d) When -B was used with -n, some line numbers in the output were one less
1314 than they should have been.
1316 (e) Added the -o (--only-matching) option.
1318 (f) If -A or -C was used with -c (count only), some lines of context were
1319 accidentally printed for the final match.
1321 (g) Added the -H (--with-filename) option.
1323 (h) The combination of options -rh failed to suppress file names for files
1324 that were found from directory arguments.
1326 (i) Added the -D (--devices) and -d (--directories) options.
1328 (j) Added the -F (--fixed-strings) option.
1330 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1332 (l) Added the --colo(u)r option.
1334 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1335 is not present by default.
1337 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1338 items such as (?R) or (?1), when the recursion could match a number of
1339 alternatives. If it matched one of the alternatives, but subsequently,
1340 outside the recursion, there was a failure, the code tried to back up into
1341 the recursion. However, because of the way PCRE is implemented, this is not
1342 possible, and the result was an incorrect result from the match.
1344 In order to prevent this happening, the specification of recursion has
1345 been changed so that all such subpatterns are automatically treated as
1346 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1348 4. I had overlooked the fact that, in some locales, there are characters for
1349 which isalpha() is true but neither isupper() nor islower() are true. In
1350 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1351 and ordfeminine) are like this. This affected the treatment of \w and \W
1352 when they appeared in character classes, but not when they appeared outside
1353 a character class. The bit map for "word" characters is now created
1354 separately from the results of isalnum() instead of just taking it from the
1355 upper, lower, and digit maps. (Plus the underscore character, of course.)
1357 5. The above bug also affected the handling of POSIX character classes such as
1358 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1359 permanent tables. Instead, the bit maps for such a class were previously
1360 created as the appropriate unions of the upper, lower, and digit bitmaps.
1361 Now they are created by subtraction from the [[:word:]] class, which has
1362 its own bitmap.
1364 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1365 It is created by subtracting the vertical space characters (\x09, \x0a,
1366 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1367 subtraction was done in the overall bitmap for a character class, meaning
1368 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1369 be recognized. This bug has been fixed.
1371 7. Patches from the folks at Google:
1373 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1374 real life, but is still worth protecting against".
1376 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1377 regular expressions".
1379 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1380 have it.
1382 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1383 "configure" and the latter not, in order to fix a problem somebody had
1384 with compiling the Arg class on HP-UX.
1386 (e) Improve the error-handling of the C++ wrapper a little bit.
1388 (f) New tests for checking recursion limiting.
1390 8. The pcre_memmove() function, which is used only if the environment does not
1391 have a standard memmove() function (and is therefore rarely compiled),
1392 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1393 returning a result (though PCRE never actually uses the result).
1395 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1396 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1397 returned instead of calling malloc() with an overflowing number that would
1398 most likely cause subsequent chaos.
1400 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1402 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1403 with this option is matched, the nmatch and pmatch options of regexec() are
1404 ignored.
1406 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1407 provided in case anyone wants to the the POSIX interface with UTF-8
1408 strings.
1410 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1411 C++ linking (needed for some HP-UX environments).
1413 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1414 (unused parameter) and in the pcre_printint() function (omitted "default"
1415 switch label when the default is to do nothing).
1417 15. Added some code to make it possible, when PCRE is compiled as a C++
1418 library, to replace subject pointers for pcre_exec() with a smart pointer
1419 class, thus making it possible to process discontinuous strings.
1421 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1422 much the same function. They were added by different people who were trying
1423 to make PCRE easy to compile on non-Unix systems. It has been suggested
1424 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1425 for compiling on Windows systems. I have therefore replaced it with
1426 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1427 defaults to "extern" for C or "extern C" for C++, which works fine on
1428 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1429 SCOPE with something explicit in config.h. In addition:
1431 (a) pcreposix.h still had just "extern" instead of either of these macros;
1432 I have replaced it with PCRE_DATA_SCOPE.
1434 (b) Functions such as _pcre_xclass(), which are internal to the library,
1435 but external in the C sense, all had PCRE_EXPORT in their definitions.
1436 This is apparently wrong for the Windows case, so I have removed it.
1437 (It makes no difference on Unix-like systems.)
1439 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1440 of recursive calls to match(). This is different to MATCH_LIMIT because
1441 that limits the total number of calls to match(), not all of which increase
1442 the depth of recursion. Limiting the recursion depth limits the amount of
1443 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1444 when PCRE is compiled, and changed at run time. A patch from Google adds
1445 this functionality to the C++ interface.
1447 18. Changes to the handling of Unicode character properties:
1449 (a) Updated the table to Unicode 4.1.0.
1451 (b) Recognize characters that are not in the table as "Cn" (undefined).
1453 (c) I revised the way the table is implemented to a much improved format
1454 which includes recognition of ranges. It now supports the ranges that
1455 are defined in UnicodeData.txt, and it also amalgamates other
1456 characters into ranges. This has reduced the number of entries in the
1457 table from around 16,000 to around 3,000, thus reducing its size
1458 considerably. I realized I did not need to use a tree structure after
1459 all - a binary chop search is just as efficient. Having reduced the
1460 number of entries, I extended their size from 6 bytes to 8 bytes to
1461 allow for more data.
1463 (d) Added support for Unicode script names via properties such as \p{Han}.
1465 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1466 matching that character.
1468 20. When matching a repeated Unicode property with a minimum greater than zero,
1469 (for example \pL{2,}), PCRE could look past the end of the subject if it
1470 reached it while seeking the minimum number of characters. This could
1471 happen only if some of the characters were more than one byte long, because
1472 there is a check for at least the minimum number of bytes.
1474 21. Refactored the implementation of \p and \P so as to be more general, to
1475 allow for more different types of property in future. This has changed the
1476 compiled form incompatibly. Anybody with saved compiled patterns that use
1477 \p or \P will have to recompile them.
1479 22. Added "Any" and "L&" to the supported property types.
1481 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1482 but give a compile time error if the value is greater than 0xff.
1484 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1485 accidentally not being installed or uninstalled.
1487 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1488 made were to insert the current release number. This seemed silly, because
1489 it made things harder for people building PCRE on systems that don't run
1490 "configure". I have turned pcre.h into a distributed file, no longer built
1491 by "configure", with the version identification directly included. There is
1492 no longer a pcre.h.in file.
1494 However, this change necessitated a change to the pcre-config script as
1495 well. It is built from pcre-config.in, and one of the substitutions was the
1496 release number. I have updated configure.ac so that ./configure now finds
1497 the release number by grepping pcre.h.
1499 26. Added the ability to run the tests under valgrind.
1502 Version 6.4 05-Sep-05
1503 ---------------------
1505 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1506 "--" to be printed when multiple files were scanned, even when none of the
1507 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1508 consider it to be a bug, and have restored the previous behaviour.
1510 2. A couple of code tidies to get rid of compiler warnings.
1512 3. The pcretest program used to cheat by referring to symbols in the library
1513 whose names begin with _pcre_. These are internal symbols that are not
1514 really supposed to be visible externally, and in some environments it is
1515 possible to suppress them. The cheating is now confined to including
1516 certain files from the library's source, which is a bit cleaner.
1518 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1519 file's purpose clearer.
1521 5. Reorganized pcre_ucp_findchar().
1524 Version 6.3 15-Aug-05
1525 ---------------------
1527 1. The file libpcre.pc.in did not have general read permission in the tarball.
1529 2. There were some problems when building without C++ support:
1531 (a) If C++ support was not built, "make install" and "make test" still
1532 tried to test it.
1534 (b) There were problems when the value of CXX was explicitly set. Some
1535 changes have been made to try to fix these, and ...
1537 (c) --disable-cpp can now be used to explicitly disable C++ support.
1539 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1540 backslash in a target when C++ was disabled. This confuses some
1541 versions of "make", apparently. Using an intermediate variable solves
1542 this. (Same for CPP_LOBJ.)
1544 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1545 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1546 necessary on certain architectures.
1548 4. Added a setting of -export-symbols-regex to the link command to remove
1549 those symbols that are exported in the C sense, but actually are local
1550 within the library, and not documented. Their names all begin with
1551 "_pcre_". This is not a perfect job, because (a) we have to except some
1552 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1553 available (and never for static libraries). I have made a note to try to
1554 find a way round (a) in the future.
1557 Version 6.2 01-Aug-05
1558 ---------------------
1560 1. There was no test for integer overflow of quantifier values. A construction
1561 such as {1111111111111111} would give undefined results. What is worse, if
1562 a minimum quantifier for a parenthesized subpattern overflowed and became
1563 negative, the calculation of the memory size went wrong. This could have
1564 led to memory overwriting.
1566 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1568 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1569 operating environments where this matters.
1571 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1572 PCRE options from within the C++ wrapper.
1574 5. Named capturing subpatterns were not being correctly counted when a pattern
1575 was compiled. This caused two problems: (a) If there were more than 100
1576 such subpatterns, the calculation of the memory needed for the whole
1577 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1578 back references of the form \12, where the number was greater than 9, were
1579 not recognized as back references, even though there were sufficient
1580 previous subpatterns.
1582 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1583 versions of gcc, e.g. 2.95.4.
1586 Version 6.1 21-Jun-05
1587 ---------------------
1589 1. There was one reference to the variable "posix" in pcretest.c that was not
1590 surrounded by "#if !defined NOPOSIX".
1592 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1593 the cross-check on the old pcre_info() function, for the benefit of the
1594 cut-down version of PCRE that is currently imported into Exim.
1596 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1597 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1598 patterns that start (?i) but I don't think that matters. The use of (?i) is
1599 just an example; this all applies to the other options as well.
1601 4. Since libtool seems to echo the compile commands it is issuing, the output
1602 from "make" can be reduced a bit by putting "@" in front of each libtool
1603 compile command.
1605 5. Patch from the folks at Google for configure.in to be a bit more thorough
1606 in checking for a suitable C++ installation before trying to compile the
1607 C++ stuff. This should fix a reported problem when a compiler was present,
1608 but no suitable headers.
1610 6. The man pages all had just "PCRE" as their title. I have changed them to
1611 be the relevant file name. I have also arranged that these names are
1612 retained in the file doc/pcre.txt, which is a concatenation in text format
1613 of all the man pages except the little individual ones for each function.
1615 7. The NON-UNIX-USE file had not been updated for the different set of source
1616 files that come with release 6. I also added a few comments about the C++
1617 wrapper.
1620 Version 6.0 07-Jun-05
1621 ---------------------
1623 1. Some minor internal re-organization to help with my DFA experiments.
1625 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1626 didn't matter for the library itself when fully configured, but did matter
1627 when compiling without UCP support, or within Exim, where the ucp files are
1628 not imported.
1630 3. Refactoring of the library code to split up the various functions into
1631 different source modules. The addition of the new DFA matching code (see
1632 below) to a single monolithic source would have made it really too
1633 unwieldy, quite apart from causing all the code to be include in a
1634 statically linked application, when only some functions are used. This is
1635 relevant even without the DFA addition now that patterns can be compiled in
1636 one application and matched in another.
1638 The downside of splitting up is that there have to be some external
1639 functions and data tables that are used internally in different modules of
1640 the library but which are not part of the API. These have all had their
1641 names changed to start with "_pcre_" so that they are unlikely to clash
1642 with other external names.
1644 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1645 a different (DFA) algorithm. Although it is slower than the original
1646 function, it does have some advantages for certain types of matching
1647 problem.
1649 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1650 including restarting after a partial match.
1652 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1653 defined when compiling for Windows was sent to me. I have put it into the
1654 code, though I have no means of testing or verifying it.
1656 7. Added the pcre_refcount() auxiliary function.
1658 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1659 match before or at the first newline in the subject string. In pcretest,
1660 the /f option on a pattern can be used to set this.
1662 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1663 would behave wrongly. This has been present in PCRE since release 4.0.
1665 10. A number of changes to the pcregrep command:
1667 (a) Refactored how -x works; insert ^(...)$ instead of setting
1668 PCRE_ANCHORED and checking the length, in preparation for adding
1669 something similar for -w.
1671 (b) Added the -w (match as a word) option.
1673 (c) Refactored the way lines are read and buffered so as to have more
1674 than one at a time available.
1676 (d) Implemented a pcregrep test script.
1678 (e) Added the -M (multiline match) option. This allows patterns to match
1679 over several lines of the subject. The buffering ensures that at least
1680 8K, or the rest of the document (whichever is the shorter) is available
1681 for matching (and similarly the previous 8K for lookbehind assertions).
1683 (f) Changed the --help output so that it now says
1685 -w, --word-regex(p)
1687 instead of two lines, one with "regex" and the other with "regexp"
1688 because that confused at least one person since the short forms are the
1689 same. (This required a bit of code, as the output is generated
1690 automatically from a table. It wasn't just a text change.)
1692 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1693 option but starts with a hyphen. Could be a pattern or a path name
1694 starting with a hyphen, for instance.
1696 (h) "-" can be given as a file name to represent stdin.
1698 (i) When file names are being printed, "(standard input)" is used for
1699 the standard input, for compatibility with GNU grep. Previously
1700 "<stdin>" was used.
1702 (j) The option --label=xxx can be used to supply a name to be used for
1703 stdin when file names are being printed. There is no short form.
1705 (k) Re-factored the options decoding logic because we are going to add
1706 two more options that take data. Such options can now be given in four
1707 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1709 (l) Added the -A, -B, and -C options for requesting that lines of context
1710 around matches be printed.
1712 (m) Added the -L option to print the names of files that do not contain
1713 any matching lines, that is, the complement of -l.
1715 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1716 continue to scan other files.
1718 (o) The -s option was incorrectly implemented. For compatibility with other
1719 greps, it now suppresses the error message for a non-existent or non-
1720 accessible file (but not the return code). There is a new option called
1721 -q that suppresses the output of matching lines, which was what -s was
1722 previously doing.
1724 (p) Added --include and --exclude options to specify files for inclusion
1725 and exclusion when recursing.
1727 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1728 Hopefully, it now does.
1730 12. Missing cast in pcre_study().
1732 13. Added an "uninstall" target to the makefile.
1734 14. Replaced "extern" in the function prototypes in Makefile.in with
1735 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1736 world, but is set differently for Windows.
1738 15. Added a second compiling function called pcre_compile2(). The only
1739 difference is that it has an extra argument, which is a pointer to an
1740 integer error code. When there is a compile-time failure, this is set
1741 non-zero, in addition to the error test pointer being set to point to an
1742 error message. The new argument may be NULL if no error number is required
1743 (but then you may as well call pcre_compile(), which is now just a
1744 wrapper). This facility is provided because some applications need a
1745 numeric error indication, but it has also enabled me to tidy up the way
1746 compile-time errors are handled in the POSIX wrapper.
1748 16. Added VPATH=.libs to the makefile; this should help when building with one
1749 prefix path and installing with another. (Or so I'm told by someone who
1750 knows more about this stuff than I do.)
1752 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1753 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1754 match everything, including newlines. This is not POSIX-compatible, but
1755 somebody wanted the feature. From pcretest it can be activated by using
1756 both the P and the s flags.
1758 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1760 19. libpcre.pc was being incorrectly installed as executable.
1762 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1763 it now also looks for '\r' so that it will work unmodified on Windows.
1765 21. Added Google's contributed C++ wrapper to the distribution.
1767 22. Added some untidy missing memory free() calls in pcretest, to keep
1768 Electric Fence happy when testing.
1772 Version 5.0 13-Sep-04
1773 ---------------------
1775 1. Internal change: literal characters are no longer packed up into items
1776 containing multiple characters in a single byte-string. Each character
1777 is now matched using a separate opcode. However, there may be more than one
1778 byte in the character in UTF-8 mode.
1780 2. The pcre_callout_block structure has two new fields: pattern_position and
1781 next_item_length. These contain the offset in the pattern to the next match
1782 item, and its length, respectively.
1784 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1785 insertion of callouts before each pattern item. Added the /C option to
1786 pcretest to make use of this.
1788 4. On the advice of a Windows user, the lines
1790 #if defined(_WIN32) || defined(WIN32)
1791 _setmode( _fileno( stdout ), 0x8000 );
1792 #endif /* defined(_WIN32) || defined(WIN32) */
1794 have been added to the source of pcretest. This apparently does useful
1795 magic in relation to line terminators.
1797 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1798 for the benefit of those environments where the "b" makes a difference.
1800 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1801 to know about it. I have put a hack into configure.in that adds in code
1802 to set GCC=yes if CC=icc. This seems to end up at a point in the
1803 generated configure script that is early enough to affect the setting of
1804 compiler options, which is what is needed, but I have no means of testing
1805 whether it really works. (The user who reported this had patched the
1806 generated configure script, which of course I cannot do.)
1808 LATER: After change 22 below (new libtool files), the configure script
1809 seems to know about icc (and also ecc). Therefore, I have commented out
1810 this hack in configure.in.
1812 7. Added support for pkg-config (2 patches were sent in).
1814 8. Negated POSIX character classes that used a combination of internal tables
1815 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1816 [[:^ascii]]. Typically, they would match almost any characters. The other
1817 POSIX classes were not broken in this way.
1819 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1820 to find the match, as PCRE was deluded into thinking that the match had to
1821 start at the start point or following a newline. The same bug applied to
1822 patterns with negative forward assertions or any backward assertions
1823 preceding ".*" at the start, unless the pattern required a fixed first
1824 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1826 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1827 starting at the last subject character, bytes beyond the end of the subject
1828 string were read.
1830 11. Renamed the variable "class" as "classbits" to make life easier for C++
1831 users. (Previously there was a macro definition, but it apparently wasn't
1832 enough.)
1834 12. Added the new field "tables" to the extra data so that tables can be passed
1835 in at exec time, or the internal tables can be re-selected. This allows
1836 a compiled regex to be saved and re-used at a later time by a different
1837 program that might have everything at different addresses.
1839 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1840 -R library as well as a -L library.
1842 14. The debugging options of pcretest (-d on the command line or D on a
1843 pattern) showed incorrect output for anything following an extended class
1844 that contained multibyte characters and which was followed by a quantifier.
1846 15. Added optional support for general category Unicode character properties
1847 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1848 support. It adds about 90K to the size of the library. The meanings of the
1849 inbuilt class escapes such as \d and \s have NOT been changed.
1851 16. Updated pcredemo.c to include calls to free() to release the memory for the
1852 compiled pattern.
1854 17. The generated file chartables.c was being created in the source directory
1855 instead of in the building directory. This caused the build to fail if the
1856 source directory was different from the building directory, and was
1857 read-only.
1859 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1860 file. No doubt somebody will tell me if they don't make sense... Also added
1861 Dan Mooney's comments about building on OpenVMS.
1863 19. Added support for partial matching via the PCRE_PARTIAL option for
1864 pcre_exec() and the \P data escape in pcretest.
1866 20. Extended pcretest with 3 new pattern features:
1868 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1869 write the compiled pattern to the file whose name is "rest-of-line".
1870 This is a straight binary dump of the data, with the saved pointer to
1871 the character tables forced to be NULL. The study data, if any, is
1872 written too. After writing, pcretest reads a new pattern.
1874 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1875 compiled pattern from the given file. There must not be any
1876 occurrences of "<" in the file name (pretty unlikely); if there are,
1877 pcretest will instead treat the initial "<" as a pattern delimiter.
1878 After reading in the pattern, pcretest goes on to read data lines as
1879 usual.
1881 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1882 and 16-bit fields in a compiled pattern, to simulate a pattern that
1883 was compiled on a host of opposite endianness.
1885 21. The pcre-exec() function can now cope with patterns that were compiled on
1886 hosts of opposite endianness, with this restriction:
1888 As for any compiled expression that is saved and used later, the tables
1889 pointer field cannot be preserved; the extra_data field in the arguments
1890 to pcre_exec() should be used to pass in a tables address if a value
1891 other than the default internal tables were used at compile time.
1893 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1894 now diagnosed as an error. Previously, most of the time, a negative number
1895 would have been treated as zero, but if in addition "ovector" was passed as
1896 NULL, a crash could occur.
1898 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1899 new versions from the libtool 1.5 distribution (the last one is a copy of
1900 a file called libtool.m4). This seems to have fixed the need to patch
1901 "configure" to support Darwin 1.3 (which I used to do). However, I still
1902 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1903 workstation).
1905 24. Changed the PCRE licence to be the more standard "BSD" licence.
1908 Version 4.5 01-Dec-03
1909 ---------------------
1911 1. There has been some re-arrangement of the code for the match() function so
1912 that it can be compiled in a version that does not call itself recursively.
1913 Instead, it keeps those local variables that need separate instances for
1914 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1915 needs to "recurse". Keeping track of where control must go is done by means
1916 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1917 hide most of the details from the main code, and operates only if
1918 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1919 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1920 operating.
1922 To make it easier for callers to provide specially tailored get/free
1923 functions for this usage, two new functions, pcre_stack_malloc, and
1924 pcre_stack_free, are used. They are always called in strict stacking order,
1925 and the size of block requested is always the same.
1927 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1928 PCRE has been compiled to use the stack or the heap for recursion. The
1929 -C option of pcretest uses this to show which version is compiled.
1931 A new data escape \S, is added to pcretest; it causes the amounts of store
1932 obtained and freed by both kinds of malloc/free at match time to be added
1933 to the output.
1935 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1936 what's available on my current Linux desktop machine.
1938 3. When matching a UTF-8 string, the test for a valid string at the start has
1939 been extended. If start_offset is not zero, PCRE now checks that it points
1940 to a byte that is the start of a UTF-8 character. If not, it returns
1941 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1942 this is necessary because there may be backward assertions in the pattern.
1943 When matching the same subject several times, it may save resources to use
1944 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1946 4. The code for checking the validity of UTF-8 strings has been tightened so
1947 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1948 containing "overlong sequences".
1950 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1951 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1952 should have been "&", but it just so happened that all the cases this let
1953 through by mistake were picked up later in the function.
1955 6. I had used a variable called "isblank" - this is a C99 function, causing
1956 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1958 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1959 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1960 so the version is shown; (c) stop "make test" echoing "./RunTest".
1962 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1964 9. The prototype for memmove() for systems that don't have it was using
1965 size_t, but the inclusion of the header that defines size_t was later. I've
1966 moved the #includes for the C headers earlier to avoid this.
1968 10. Added some adjustments to the code to make it easier to compiler on certain
1969 special systems:
1971 (a) Some "const" qualifiers were missing.
1972 (b) Added the macro EXPORT before all exported functions; by default this
1973 is defined to be empty.
1974 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1975 that it reads its output file name as an argument instead of writing
1976 to the standard output and assuming this can be redirected.
1978 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1979 class containing characters with values greater than 255, PCRE compilation
1980 went into a loop.
1982 12. A recursive reference to a subpattern that was within another subpattern
1983 that had a minimum quantifier of zero caused PCRE to crash. For example,
1984 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1985 recursion. If the recursively-called subpattern itself had a zero repeat,
1986 that was OK.
1988 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1989 buffer into which it was copied (for escape processing) was still set at
1990 1024, so long lines caused crashes.
1992 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1993 "internal error: code overflow...". This applied to any character class
1994 that was followed by a possessive quantifier.
1996 15. Modified the Makefile to add libpcre.la as a prerequisite for
1997 libpcreposix.la because I was told this is needed for a parallel build to
1998 work.
2000 16. If a pattern that contained .* following optional items at the start was
2001 studied, the wrong optimizing data was generated, leading to matching
2002 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2003 matching string must start with a or b or c. The correct conclusion for
2004 this pattern is that a match can start with any character.
2007 Version 4.4 13-Aug-03
2008 ---------------------
2010 1. In UTF-8 mode, a character class containing characters with values between
2011 127 and 255 was not handled correctly if the compiled pattern was studied.
2012 In fixing this, I have also improved the studying algorithm for such
2013 classes (slightly).
2015 2. Three internal functions had redundant arguments passed to them. Removal
2016 might give a very teeny performance improvement.
2018 3. Documentation bug: the value of the capture_top field in a callout is *one
2019 more than* the number of the hightest numbered captured substring.
2021 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2022 in incorrectly linking with a previously installed version. They now link
2023 explicitly with libpcre.la.
2025 5. configure.in no longer needs to recognize Cygwin specially.
2027 6. A problem in pcre.in for Windows platforms is fixed.
2029 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2030 pcretest, it used to include the size of the study block as part of its
2031 output. Unfortunately, the structure contains a field that has a different
2032 size on different hardware architectures. This meant that the tests that
2033 showed this size failed. As the block is currently always of a fixed size,
2034 this information isn't actually particularly useful in pcretest output, so
2035 I have just removed it.
2037 8. Three pre-processor statements accidentally did not start in column 1.
2038 Sadly, there are *still* compilers around that complain, even though
2039 standard C has not required this for well over a decade. Sigh.
2041 9. In pcretest, the code for checking callouts passed small integers in the
2042 callout_data field, which is a void * field. However, some picky compilers
2043 complained about the casts involved for this on 64-bit systems. Now
2044 pcretest passes the address of the small integer instead, which should get
2045 rid of the warnings.
2047 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2048 both compile and run time, and gives an error if an invalid UTF-8 sequence
2049 is found. There is a option for disabling this check in cases where the
2050 string is known to be correct and/or the maximum performance is wanted.
2052 11. In response to a bug report, I changed one line in Makefile.in from
2054 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2055 to
2056 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2058 to look similar to other lines, but I have no way of telling whether this
2059 is the right thing to do, as I do not use Windows. No doubt I'll get told
2060 if it's wrong...
2063 Version 4.3 21-May-03
2064 ---------------------
2066 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2067 Makefile.
2069 2. Some refactoring to improve the quality of the code:
2071 (i) The utf8_table... variables are now declared "const".
2073 (ii) The code for \cx, which used the "case flipping" table to upper case
2074 lower case letters, now just substracts 32. This is ASCII-specific,
2075 but the whole concept of \cx is ASCII-specific, so it seems
2076 reasonable.
2078 (iii) PCRE was using its character types table to recognize decimal and
2079 hexadecimal digits in the pattern. This is silly, because it handles
2080 only 0-9, a-f, and A-F, but the character types table is locale-
2081 specific, which means strange things might happen. A private
2082 table is now used for this - though it costs 256 bytes, a table is
2083 much faster than multiple explicit tests. Of course, the standard
2084 character types table is still used for matching digits in subject
2085 strings against \d.
2087 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2088 ending in _t are). So I've renamed it as ESC_tee.
2090 3. The first argument for regexec() in the POSIX wrapper should have been
2091 defined as "const".
2093 4. Changed pcretest to use malloc() for its buffers so that they can be
2094 Electric Fenced for debugging.
2096 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2097 to read one or more bytes before the start of the subject string. Often this
2098 had no effect on PCRE's behaviour, but in some circumstances it could
2099 provoke a segmentation fault.
2101 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2102 to try to read one or more bytes before the start of the subject string.
2104 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2105 UTF-8 support could misbehave in various ways if the subject string
2106 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2107 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2108 back over UTF-8 characters.)
2111 Version 4.2 14-Apr-03
2112 ---------------------
2114 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2116 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2117 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2118 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2119 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2120 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2122 Note: automatic setting of the BUILD variables is not yet working
2123 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2124 compile-time but not at link-time
2125 [LINK]: use for linking executables only
2126 make different versions for Windows and non-Windows
2127 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2128 libraries
2129 [LINK_FOR_BUILD]: new variable
2130 [OBJEXT]: use throughout
2131 [EXEEXT]: use throughout
2132 <winshared>: new target
2133 <wininstall>: new target
2134 <dftables.o>: use native compiler
2135 <dftables>: use native linker
2136 <install>: handle Windows platform correctly
2137 <clean>: ditto
2138 <check>: ditto
2139 copy DLL to top builddir before testing
2141 As part of these changes, -no-undefined was removed again. This was reported
2142 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2143 in any case.
2145 3. Some tidies to get rid of compiler warnings:
2147 . In the match_data structure, match_limit was an unsigned long int, whereas
2148 match_call_count was an int. I've made them both unsigned long ints.
2150 . In pcretest the fact that a const uschar * doesn't automatically cast to
2151 a void * provoked a warning.
2153 . Turning on some more compiler warnings threw up some "shadow" variables
2154 and a few more missing casts.
2156 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2157 option, a class that contained a single character with a value between 128
2158 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2160 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2161 option, a class that contained several characters, but with at least one
2162 whose value was between 128 and 255 caused PCRE to crash.
2165 Version 4.1 12-Mar-03
2166 ---------------------
2168 1. Compiling with gcc -pedantic found a couple of places where casts were
2169 needed, and a string in dftables.c that was longer than standard compilers are
2170 required to support.
2172 2. Compiling with Sun's compiler found a few more places where the code could
2173 be tidied up in order to avoid warnings.
2175 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2176 first of these names is deprecated in the latest Autoconf in favour of the name
2177 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2178 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2179 analogy I have changed it to CFLAGS_FOR_BUILD.
2181 4. Added -no-undefined to the linking command in the Makefile, because this is
2182 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2183 linking step for the pcreposix library.
2185 5. PCRE was failing to diagnose the case of two named groups with the same
2186 name.
2188 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2189 literal character that is needed in the subject for a match, and scans along to
2190 ensure that it is present before embarking on the full matching process. This
2191 saves time in cases of nested unlimited repeats that are never going to match.
2192 Problem: the scan can take a lot of time if the subject is very long (e.g.
2193 megabytes), thus penalizing straightforward matches. It is now done only if the
2194 amount of subject to be scanned is less than 1000 bytes.
2196 7. A lesser problem with the same optimization is that it was recording the
2197 first character of an anchored pattern as "needed", thus provoking a search
2198 right along the subject, even when the first match of the pattern was going to
2199 fail. The "needed" character is now not set for anchored patterns, unless it
2200 follows something in the pattern that is of non-fixed length. Thus, it still
2201 fulfils its original purpose of finding quick non-matches in cases of nested
2202 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2205 Version 4.0 17-Feb-03
2206 ---------------------
2208 1. If a comment in an extended regex that started immediately after a meta-item
2209 extended to the end of string, PCRE compiled incorrect data. This could lead to
2210 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2212 2. Moved to autoconf 2.53 and libtool 1.4.2.
2214 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2215 the special perltest8 script is no longer needed - all the tests can be run
2216 from a single perltest script.
2218 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2219 by \s. It has now been removed in PCRE. This means it isn't recognized as
2220 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2221 class [:space:] *does* include VT, thereby creating a mess.
2223 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2224 space and tab.
2226 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2227 its new features into the main test script, reducing the number of scripts.
2229 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2230 were backward compatible, and made the (?i) apply to the whole pattern, as if
2231 /i were given. Now it behaves more logically, and applies the option setting
2232 only to what follows. PCRE has been changed to follow suit. However, if it
2233 finds options settings right at the start of the pattern, it extracts them into
2234 the global options, as before. Thus, they show up in the info data.
2236 8. Added support for the \Q...\E escape sequence. Characters in between are
2237 treated as literals. This is slightly different from Perl in that $ and @ are
2238 also handled as literals inside the quotes. In Perl, they will cause variable
2239 interpolation. Note the following examples:
2241 Pattern PCRE matches Perl matches
2243 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2244 \Qabc\$xyz\E abc\$xyz abc\$xyz
2245 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2247 For compatibility with Perl, \Q...\E sequences are recognized inside character
2248 classes as well as outside them.
2250 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2251 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2252 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2253 signed/unsigned warnings.
2255 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2256 option for pcretest, so I've replaced it by a simple function that does just
2257 that job.
2259 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2260 "pcregrep -".
2262 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2263 Java package. This provides some syntactic sugar for simple cases of what my
2264 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2265 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2266 item, you can use this simplified notation. Note that only makes sense with
2267 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2268 greediness, whatever the setting of the PCRE_UNGREEDY option.
2270 13. A change of greediness default within a pattern was not taking effect at
2271 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2272 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2273 was abstracted outside.
2275 14. PCRE now supports the \G assertion. It is true when the current matching
2276 position is at the start point of the match. This differs from \A when the
2277 starting offset is non-zero. Used with the /g option of pcretest (or similar
2278 code), it works in the same way as it does for Perl's /g option. If all
2279 alternatives of a regex begin with \G, the expression is anchored to the start
2280 match position, and the "anchored" flag is set in the compiled expression.
2282 15. Some bugs concerning the handling of certain option changes within patterns
2283 have been fixed. These applied to options other than (?ims). For example,
2284 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2285 been the other way round. Some of this was related to change 7 above.
2287 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2288 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2289 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2290 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2292 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2293 mode. Unlike ".", it always matches newline, whatever the setting of
2294 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2295 assertions. Perl allows it, but it doesn't (in general) work because it can't
2296 calculate the length of the lookbehind. At least, that's the case for Perl
2297 5.8.0 - I've been told they are going to document that it doesn't work in
2298 future.
2300 18. Added an error diagnosis for escapes that PCRE does not support: these are
2301 \L, \l, \N, \P, \p, \U, \u, and \X.
2303 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2304 reading past the end of the pattern in cases such as /[abcd/.
2306 20. PCRE was getting more memory than necessary for patterns with classes that
2307 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2309 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2310 compiling PCRE for use with Virtual Pascal.
2312 22. Small fix to the Makefile to make it work properly if the build is done
2313 outside the source tree.
2315 23. Added a new extension: a condition to go with recursion. If a conditional
2316 subpattern starts with (?(R) the "true" branch is used if recursion has
2317 happened, whereas the "false" branch is used only at the top level.
2319 24. When there was a very long string of literal characters (over 255 bytes
2320 without UTF support, over 250 bytes with UTF support), the computation of how
2321 much memory was required could be incorrect, leading to segfaults or other
2322 strange effects.
2324 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2325 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2326 there was a subsequent back reference to those brackets. This meant that, for
2327 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2328 possible to check for precisely this case. All we can do is abandon the
2329 optimization if .* occurs inside capturing brackets when there are any back
2330 references whatsoever. (See below for a better fix that came later.)
2332 26. The handling of the optimization for finding the first character of a
2333 non-anchored pattern, and for finding a character that is required later in the
2334 match were failing in some cases. This didn't break the matching; it just
2335 failed to optimize when it could. The way this is done has been re-implemented.
2337 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2339 28. Added a new feature that provides some of the functionality that Perl
2340 provides with (?{...}). The facility is termed a "callout". The way it is done
2341 in PCRE is for the caller to provide an optional function, by setting
2342 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2343 global variable. By default it is unset, which disables all calling out. To get
2344 the function called, the regex must include (?C) at appropriate points. This
2345 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2346 This provides a means of identifying different callout points. When PCRE
2347 reaches such a point in the regex, if pcre_callout has been set, the external
2348 function is called. It is provided with data in a structure called
2349 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2350 matching continues; if it returns a non-zero value, the match at the current
2351 point fails. However, backtracking will occur if possible. [This was changed
2352 later and other features added - see item 49 below.]
2354 29. pcretest is upgraded to test the callout functionality. It provides a
2355 callout function that displays information. By default, it shows the start of
2356 the match and the current position in the text. There are some new data escapes
2357 to vary what happens:
2359 \C+ in addition, show current contents of captured substrings
2360 \C- do not supply a callout function
2361 \C!n return 1 when callout number n is reached
2362 \C!n!m return 1 when callout number n is reached for the mth time
2364 30. If pcregrep was called with the -l option and just a single file name, it
2365 output "<stdin>" if a match was found, instead of the file name.
2367 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2368 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2369 pcre_exec(). This saves a malloc/free per call. The default value of
2370 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2371 when configuring.
2373 32. The default maximum size of a compiled pattern is 64K. There have been a
2374 few cases of people hitting this limit. The code now uses macros to handle the
2375 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2376 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2377 configuring. Tests 2 and 5 work only with 2-byte links because they output
2378 debugging information about compiled patterns.
2380 33. Internal code re-arrangements:
2382 (a) Moved the debugging function for printing out a compiled regex into
2383 its own source file (printint.c) and used #include to pull it into
2384 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2385 separate copies.
2387 (b) Defined the list of op-code names for debugging as a macro in
2388 internal.h so that it is next to the definition of the opcodes.
2390 (c) Defined a table of op-code lengths for simpler skipping along compiled
2391 code. This is again a macro in internal.h so that it is next to the
2392 definition of the opcodes.
2394 34. Added support for recursive calls to individual subpatterns, along the
2395 lines of Robin Houston's patch (but implemented somewhat differently).
2397 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2398 allow it to read and process whole directories in Win32. This code was
2399 contributed by Lionel Fourquaux; it has not been tested by me.
2401 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2402 used to name a group. Names consist of alphanumerics and underscores, and must
2403 be unique. Back references use the syntax (?P=name) and recursive calls use
2404 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2405 numbers. The function pcre_fullinfo() can be used after compilation to extract
2406 a name/number map. There are three relevant calls:
2408 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2409 PCRE_INFO_NAMECOUNT yields the number of entries
2410 PCRE_INFO_NAMETABLE yields a pointer to the map.
2412 The map is a vector of fixed-size entries. The size of each entry depends on
2413 the length of the longest name used. The first two bytes of each entry are the
2414 group number, most significant byte first. There follows the corresponding
2415 name, zero terminated. The names are in alphabetical order.
2417 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2418 case instead of 255. Making it the same both with and without UTF-8 support
2419 means that the same test output works with both.
2421 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2422 calling malloc() with a zero argument.
2424 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2425 optimization. I've improved things by keeping a bitmap of backreferences with
2426 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2427 fact referenced, the optimization can be applied. It is unlikely that a
2428 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2429 the match to follow \n) will appear inside brackets with a number greater than
2430 31, but if it does, any back reference > 31 suppresses the optimization.
2432 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2433 of disabling numbered capturing parentheses. Any opening parenthesis that is
2434 not followed by ? behaves as if it were followed by ?: but named parentheses
2435 can still be used for capturing (and they will acquire numbers in the usual
2436 way).
2438 41. Redesigned the return codes from the match() function into yes/no/error so
2439 that errors can be passed back from deep inside the nested calls. A malloc
2440 failure while inside a recursive subpattern call now causes the
2441 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2443 42. It is now possible to set a limit on the number of times the match()
2444 function is called in a call to pcre_exec(). This facility makes it possible to
2445 limit the amount of recursion and backtracking, though not in a directly
2446 obvious way, because the match() function is used in a number of different
2447 circumstances. The count starts from zero for each position in the subject
2448 string (for non-anchored patterns). The default limit is, for compatibility, a
2449 large number, namely 10 000 000. You can change this in two ways:
2451 (a) When configuring PCRE before making, you can use --with-match-limit=n
2452 to set a default value for the compiled library.
2454 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2455 a different value is set. See 45 below.
2457 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2459 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2460 of things that can be changed at compile time. The first argument specifies
2461 what is wanted and the second points to where the information is to be placed.
2462 The current list of available information is:
2466 The output is an integer that is set to one if UTF-8 support is available;
2467 otherwise it is set to zero.
2471 The output is an integer that it set to the value of the code that is used for
2472 newline. It is either LF (10) or CR (13).
2476 The output is an integer that contains the number of bytes used for internal
2477 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2481 The output is an integer that contains the threshold above which the POSIX
2482 interface uses malloc() for output vectors. See item 31 above.
2486 The output is an unsigned integer that contains the default limit of the number
2487 of match() calls in a pcre_exec() execution. See 42 above.
2489 44. pcretest has been upgraded by the addition of the -C option. This causes it
2490 to extract all the available output from the new pcre_config() function, and to
2491 output it. The program then exits immediately.
2493 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2494 order to support additional features. One way would have been to define
2495 pcre_exec2() (for example) with extra arguments, but this would not have been
2496 extensible, and would also have required all calls to the original function to
2497 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2498 is used for passing in "extra" data from pcre_study().
2500 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2501 contains the following fields:
2503 flags a bitmap indicating which of the following fields are set
2504 study_data opaque data from pcre_study()
2505 match_limit a way of specifying a limit on match() calls for a specific
2506 call to pcre_exec()
2507 callout_data data for callouts (see 49 below)
2509 The flag bits are also defined in pcre.h, and are
2515 The pcre_study() function now returns one of these new pcre_extra blocks, with
2516 the actual study data pointed to by the study_data field, and the
2517 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2518 before. That is, this change is entirely upwards-compatible and requires no
2519 change to existing code.
2521 If you want to pass in additional data to pcre_exec(), you can either place it
2522 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2523 block.
2525 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2526 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2527 times with different match limits, until it finds the minimum value needed for
2528 pcre_exec() to complete. The value is then output. This can be instructive; for
2529 most simple matches the number is quite small, but for pathological cases it
2530 gets very large very quickly.
2532 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2533 returns the size of the data block pointed to by the study_data field in a
2534 pcre_extra block, that is, the value that was passed as the argument to
2535 pcre_malloc() when PCRE was getting memory in which to place the information
2536 created by pcre_study(). The fourth argument should point to a size_t variable.
2537 pcretest has been extended so that this information is shown after a successful
2538 pcre_study() call when information about the compiled regex is being displayed.
2540 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2541 because what follows is always an absolute path. (Later: it turns out that this
2542 is more than cosmetic for MinGW, because it doesn't like empty path
2543 components.)
2545 49. Some changes have been made to the callout feature (see 28 above):
2547 (i) A callout function now has three choices for what it returns:
2549 0 => success, carry on matching
2550 > 0 => failure at this point, but backtrack if possible
2551 < 0 => serious error, return this value from pcre_exec()
2553 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2554 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2555 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2556 use by callout functions. It will never be used by PCRE itself.
2558 (ii) The pcre_extra structure (see 45 above) has a void * field called
2559 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2560 pcre_callout_block structure has a field of the same name. The contents of
2561 the field passed in the pcre_extra structure are passed to the callout
2562 function in the corresponding field in the callout block. This makes it
2563 easier to use the same callout-containing regex from multiple threads. For
2564 testing, the pcretest program has a new data escape
2566 \C*n pass the number n (may be negative) as callout_data
2568 If the callout function in pcretest receives a non-zero value as
2569 callout_data, it returns that value.
2571 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2572 there were some redundant $(CFLAGS) in commands that are now specified as
2573 $(LINK), which already includes $(CFLAGS).
2575 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2576 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2577 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2578 one-byte characters throughout. Note that case-insensitive matching applies
2579 only to characters whose values are less than 256. PCRE doesn't support the
2580 notion of cases for higher-valued characters.
2582 (i) A character class whose characters are all within 0-255 is handled as
2583 a bit map, and the map is inverted for negative classes. Previously, a
2584 character > 255 always failed to match such a class; however it should
2585 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2587 (ii) A negated character class with a single character < 255 is coded as
2588 "not this character" (OP_NOT). This wasn't working properly when the test
2589 character was multibyte, either singly or repeated.
2591 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2592 mode, for example: \x{100}{2,3}.
2594 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2595 singly or repeated) now correctly test multibyte characters. However,
2596 PCRE doesn't recognize any characters with values greater than 255 as
2597 digits, spaces, or word characters. Such characters always match \D, \S,
2598 and \W, and never match \d, \s, or \w.
2600 (v) Classes may now contain characters and character ranges with values
2601 greater than 255. For example: [ab\x{100}-\x{400}].
2603 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2604 PCRE in UTF-8 mode.
2606 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2607 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2608 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2609 value.)
2611 53. The single man page has become too large. I have therefore split it up into
2612 a number of separate man pages. These also give rise to individual HTML pages;
2613 these are now put in a separate directory, and there is an index.html page that
2614 lists them all. Some hyperlinking between the pages has been installed.
2616 54. Added convenience functions for handling named capturing parentheses.
2618 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2619 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2620 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2621 are faulted.
2623 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2624 calling configure. These values are used when compiling the dftables.c program
2625 which is run to generate the source of the default character tables. They
2626 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2627 you will need to set these values.
2629 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2632 Version 3.9 02-Jan-02
2633 ---------------------
2635 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2637 2. If --disable-static was given, the building process failed when trying to
2638 build pcretest and pcregrep. (For some reason it was using libtool to compile
2639 them, which is not right, as they aren't part of the library.)
2642 Version 3.8 18-Dec-01
2643 ---------------------
2645 1. The experimental UTF-8 code was completely screwed up. It was packing the
2646 bytes in the wrong order. How dumb can you get?
2649 Version 3.7 29-Oct-01
2650 ---------------------
2652 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2653 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2654 this didn't happen under Solaris 8, where I normally test things.
2656 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2657 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2658 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2659 this fix an hour or so after the initial 3.7 release.)
2662 Version 3.6 23-Oct-01
2663 ---------------------
2665 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2666 offsets passed as NULL with zero offset count.
2668 2. The config.guess and config.sub files had not been updated when I moved to
2669 the latest autoconf.
2672 Version 3.5 15-Aug-01
2673 ---------------------
2675 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2676 had been forgotten.
2678 2. By using declared but undefined structures, we can avoid using "void"
2679 definitions in pcre.h while keeping the internal definitions of the structures
2680 private.
2682 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2683 user point of view, this means that both static and shared libraries are built
2684 by default, but this can be individually controlled. More of the work of
2685 handling this static/shared cases is now inside libtool instead of PCRE's make
2686 file.
2688 4. The pcretest utility is now installed along with pcregrep because it is
2689 useful for users (to test regexs) and by doing this, it automatically gets
2690 relinked by libtool. The documentation has been turned into a man page, so
2691 there are now .1, .txt, and .html versions in /doc.
2693 5. Upgrades to pcregrep:
2694 (i) Added long-form option names like gnu grep.
2695 (ii) Added --help to list all options with an explanatory phrase.
2696 (iii) Added -r, --recursive to recurse into sub-directories.
2697 (iv) Added -f, --file to read patterns from a file.
2699 6. pcre_exec() was referring to its "code" argument before testing that
2700 argument for NULL (and giving an error if it was NULL).
2702 7. Upgraded Makefile.in to allow for compiling in a different directory from
2703 the source directory.
2705 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2706 options bits, the pointer it was passed was to an int instead of to an unsigned
2707 long int. This mattered only on 64-bit systems.
2709 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2710 generated) instead of pcre.in, which it its source. Also made the same change
2711 in several of the .c files.
2713 10. A new release of gcc defines printf() as a macro, which broke pcretest
2714 because it had an ifdef in the middle of a string argument for printf(). Fixed
2715 by using separate calls to printf().
2717 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2718 script, to force use of CR or LF instead of \n in the source. On non-Unix
2719 systems, the value can be set in config.h.
2721 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2722 absolute limit. Changed the text of the error message to make this clear, and
2723 likewise updated the man page.
2725 13. The limit of 99 on the number of capturing subpatterns has been removed.
2726 The new limit is 65535, which I hope will not be a "real" limit.
2729 Version 3.4 22-Aug-00
2730 ---------------------
2732 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2734 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2737 Version 3.3 01-Aug-00
2738 ---------------------
2740 1. If an octal character was given, but the value was greater than \377, it
2741 was not getting masked to the least significant bits, as documented. This could
2742 lead to crashes in some systems.
2744 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2745 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2747 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2748 These just pass their arguments on to (pcre_free)(), but they are provided
2749 because some uses of PCRE bind it to non-C systems that can call its functions,
2750 but cannot call free() or pcre_free() directly.
2752 4. Add "make test" as a synonym for "make check". Corrected some comments in
2753 the Makefile.
2755 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2756 Makefile.
2758 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2759 command called pgrep for grepping around the active processes.
2761 7. Added the beginnings of support for UTF-8 character strings.
2763 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2764 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2765 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2766 out for the ar command.)
2769 Version 3.2 12-May-00
2770 ---------------------
2772 This is purely a bug fixing release.
2774 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2775 of ZA. This was just one example of several cases that could provoke this bug,
2776 which was introduced by change 9 of version 2.00. The code for breaking
2777 infinite loops after an iteration that matches an empty string was't working
2778 correctly.
2780 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2781 when matched against abbab (for example). After matching an empty string, it
2782 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2783 caused it to match further down the string than it should.
2785 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2786 was there because it doesn't seem to be needed, and it causes trouble on some
2787 systems, as it is not a Standard C header. It has been removed.
2789 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2790 were reported on the Macintosh. The changes were from
2792 while ((c = *(++ptr)) != 0 && c != '\n');
2793 to
2794 while ((c = *(++ptr)) != 0 && c != '\n') ;
2796 Totally extraordinary, but if that's what it takes...
2798 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2799 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2800 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2801 assumes the way PCRE uses memmove() (always moving upwards).
2803 6. PCRE is being used in one environment where strchr() is not available. There
2804 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2805 faster code anyway.
2808 Version 3.1 09-Feb-00
2809 ---------------------
2811 The only change in this release is the fixing of some bugs in Makefile.in for
2812 the "install" target:
2814 (1) It was failing to install pcreposix.h.
2816 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2819 Version 3.0 01-Feb-00
2820 ---------------------
2822 1. Add support for the /+ modifier to perltest (to output $` like it does in
2823 pcretest).
2825 2. Add support for the /g modifier to perltest.
2827 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2828 matches null strings.
2830 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2831 pattern. Perl treats empty patterns specially - it reuses the most recent
2832 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2833 effect.
2835 5. The POSIX interface was broken in that it was just handing over the POSIX
2836 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2837 required a bigger vector, with some working space on the end. This means that
2838 the POSIX wrapper now has to get and free some memory, and copy the results.
2840 6. Added some simple autoconf support, placing the test data and the
2841 documentation in separate directories, re-organizing some of the
2842 information files, and making it build pcre-config (a GNU standard). Also added
2843 libtool support for building PCRE as a shared library, which is now the
2844 default.
2846 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2847 09 are not valid octal constants. Single digits will be used for minor values
2848 less than 10.
2850 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2851 existing programs that set these in the POSIX interface can use PCRE without
2852 modification.
2854 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2855 return all that pcre_info() returns, plus additional data. The pcre_info()
2856 function is retained for compatibility, but is considered to be obsolete.
2858 10. Added experimental recursion feature (?R) to handle one common case that
2859 Perl 5.6 will be able to do with (?p{...}).
2861 11. Added support for POSIX character classes like [:alpha:], which Perl is
2862 adopting.
2865 Version 2.08 31-Aug-99
2866 ----------------------
2868 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2869 trying to match at the startoffset position, but instead was moving forward to
2870 the next newline as if a previous match had failed.
2872 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2873 and could get into a loop if a null string was matched other than at the start
2874 of the subject.
2876 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2877 be distinguished at compile time, and for completeness also added PCRE_DATE.
2879 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2880 in GnuWin32 environments.
2883 Version 2.07 29-Jul-99
2884 ----------------------
2886 1. The documentation is now supplied in plain text form and HTML as well as in
2887 the form of man page sources.
2889 2. C++ compilers don't like assigning (void *) values to other pointer types.
2890 In particular this affects malloc(). Although there is no problem in Standard
2891 C, I've put in casts to keep C++ compilers happy.
2893 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2894 should be (const char *).
2896 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2897 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2898 However, I haven't made this a standard facility. The documentation doesn't
2899 mention it, and the Makefile doesn't support it.
2901 5. The Makefile now contains an "install" target, with editable destinations at
2902 the top of the file. The pcretest program is not installed.
2904 6. pgrep -V now gives the PCRE version number and date.
2906 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2907 causing the entire string to be ignored, instead of just the last character.
2909 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2910 non-matching string, it can take a very, very long time, even for strings of
2911 quite modest length, because of the nested recursion. PCRE now does better in
2912 some of these cases. It does this by remembering the last required literal
2913 character in the pattern, and pre-searching the subject to ensure it is present
2914 before running the real match. In other words, it applies a heuristic to detect
2915 some types of certain failure quickly, and in the above example, if presented
2916 with a string that has no trailing " it gives "no match" very quickly.
2918 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2919 other alternatives are tried instead.
2922 Version 2.06 09-Jun-99
2923 ----------------------
2925 1. Change pcretest's output for amount of store used to show just the code
2926 space, because the remainder (the data block) varies in size between 32-bit and
2927 64-bit systems.
2929 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2930 start matching at. This allows lookbehinds to work when searching for multiple
2931 occurrences in a string.
2933 3. Added additional options to pcretest for testing multiple occurrences:
2935 /+ outputs the rest of the string that follows a match
2936 /g loops for multiple occurrences, using the new startoffset argument
2937 /G loops for multiple occurrences by passing an incremented pointer
2939 4. PCRE wasn't doing the "first character" optimization for patterns starting
2940 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2941 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2942 the letter 'x'. On long subject strings, this gives a significant speed-up.
2945 Version 2.05 21-Apr-99
2946 ----------------------
2948 1. Changed the type of magic_number from int to long int so that it works
2949 properly on 16-bit systems.
2951 2. Fixed a bug which caused patterns starting with .* not to work correctly
2952 when the subject string contained newline characters. PCRE was assuming
2953 anchoring for such patterns in all cases, which is not correct because .* will
2954 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2955 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2956 must be retried after every newline in the subject.
2959 Version 2.04 18-Feb-99
2960 ----------------------
2962 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2963 computation of the store needed to hold the pattern was incorrect (too large).
2964 If such patterns were nested a few deep, this could multiply and become a real
2965 problem.
2967 2. Added /M option to pcretest to show the memory requirement of a specific
2968 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2970 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2971 compiled in such a way that the backtracking after subsequent failure was
2972 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2973 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2976 Version 2.03 02-Feb-99
2977 ----------------------
2979 1. Fixed typo and small mistake in man page.
2981 2. Added 4th condition (GPL supersedes if conflict) and created separate
2982 LICENCE file containing the conditions.
2984 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2985 Perl, that is the internal \ allows the delimiter to be included in the
2986 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2987 the final delimiter, add \ to the end of the pattern (to test the error).
2989 4. Added the convenience functions for extracting substrings after a successful
2990 match. Updated pcretest to make it able to test these functions.
2993 Version 2.02 14-Jan-99
2994 ----------------------
2996 1. Initialized the working variables associated with each extraction so that
2997 their saving and restoring doesn't refer to uninitialized store.
2999 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3000 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3001 fix the problem.
3003 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3004 calls, and wasn't printing the correct value for compiling calls. Increased the
3005 default value of LOOPREPEAT, and the number of significant figures in the
3006 times.
3008 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3010 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3011 a building problem on Windows NT with a FAT file system.
3014 Version 2.01 21-Oct-98
3015 ----------------------
3017 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3018 to character tables built by pcre_maketables() in the current locale. If NULL
3019 is passed, the default tables are used.
3022 Version 2.00 24-Sep-98
3023 ----------------------
3025 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3026 it any more.
3028 2. Allow quantification of (?>) groups, and make it work correctly.
3030 3. The first character computation wasn't working for (?>) groups.
3032 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3033 end of the subject) and add 5.005's \z, which really does match only at the
3034 very end of the subject.
3036 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3038 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3039 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3040 localized options. All options to pcre_study() were also removed.
3042 7. Add other new features from 5.005:
3044 $(?<= positive lookbehind
3045 $(?<! negative lookbehind
3046 (?imsx-imsx) added the unsetting capability
3047 such a setting is global if at outer level; local otherwise
3048 (?imsx-imsx:) non-capturing groups with option setting
3049 (?(cond)re|re) conditional pattern matching
3051 A backreference to itself in a repeated group matches the previous
3052 captured string.
3054 8. General tidying up of studying (both automatic and via "study")
3055 consequential on the addition of new assertions.
3057 9. As in 5.005, unlimited repeated groups that could match an empty substring
3058 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3059 runtime if any iteration does actually match an empty substring.
3061 10. Include the RunTest script in the distribution.
3063 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3064 discrepancies, some of which were old and were also with respect to 5.004. They
3065 have now been fixed.
3068 Version 1.09 28-Apr-98
3069 ----------------------
3071 1. A negated single character class followed by a quantifier with a minimum
3072 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3073 program crashes, or just wrong answers. This did not apply to negated classes
3074 containing more than one character, or to minima other than one.
3077 Version 1.08 27-Mar-98
3078 ----------------------
3080 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3082 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3083 latter must appear before anything that relies on it in the pattern.
3086 Version 1.07 16-Feb-98
3087 ----------------------
3089 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3090 repeat of a potentially empty string).
3093 Version 1.06 23-Jan-98
3094 ----------------------
3096 1. Added Markus Oberhumer's little patches for C++.
3098 2. Literal strings longer than 255 characters were broken.
3101 Version 1.05 23-Dec-97
3102 ----------------------
3104 1. Negated character classes containing more than one character were failing if
3105 PCRE_CASELESS was set at run time.
3108 Version 1.04 19-Dec-97
3109 ----------------------
3111 1. Corrected the man page, where some "const" qualifiers had been omitted.
3113 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3114 input syntax.
3116 3. Fixed memory leak which occurred when a regex with back references was
3117 matched with an offsets vector that wasn't big enough. The temporary memory
3118 that is used in this case wasn't being freed if the match failed.
3120 4. Tidied pcretest to ensure it frees memory that it gets.
3122 5. Temporary memory was being obtained in the case where the passed offsets
3123 vector was exactly big enough.
3125 6. Corrected definition of offsetof() from change 5 below.
3127 7. I had screwed up change 6 below and broken the rules for the use of
3128 setjmp(). Now fixed.
3131 Version 1.03 18-Dec-97
3132 ----------------------
3134 1. A erroneous regex with a missing opening parenthesis was correctly
3135 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3136 on some systems.
3138 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3139 it was reported that one broken compiler failed on the former because "code" is
3140 also an independent variable.
3142 3. The erroneous regex a[]b caused an array overrun reference.
3144 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3145 fail on data ending with that character. (It was going on too far, and checking
3146 the next character, typically a binary zero.) This was specific to the
3147 optimized code for single-character negative classes.
3149 5. Added a contributed patch from the TIN world which does the following:
3151 + Add an undef for memmove, in case the the system defines a macro for it.
3153 + Add a definition of offsetof(), in case there isn't one. (I don't know
3154 the reason behind this - offsetof() is part of the ANSI standard - but
3155 it does no harm).
3157 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3158 most of the places where whitespace preceded '#'. I have given up and
3159 allowed the remaining 2 cases to be at the margin.
3161 + Rename some variables in pcre to eliminate shadowing. This seems very
3162 pedantic, but does no harm, of course.
3164 6. Moved the call to setjmp() into its own function, to get rid of warnings
3165 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3167 7. Constructs such as \d{8,} were compiling into the equivalent of
3168 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3169 outcome, but in this particular case used more store than had been allocated,
3170 which caused the bug to be discovered because it threw up an internal error.
3172 8. The debugging code in both pcre and pcretest for outputting the compiled
3173 form of a regex was going wrong in the case of back references followed by
3174 curly-bracketed repeats.
3177 Version 1.02 12-Dec-97
3178 ----------------------
3180 1. Typos in pcre.3 and comments in the source fixed.
3182 2. Applied a contributed patch to get rid of places where it used to remove
3183 'const' from variables, and fixed some signed/unsigned and uninitialized
3184 variable warnings.
3186 3. Added the "runtest" target to Makefile.
3188 4. Set default compiler flag to -O2 rather than just -O.
3191 Version 1.01 19-Nov-97
3192 ----------------------
3194 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3195 like /([ab]*)*/, that is, for classes with more than one character in them.
3197 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3198 as /((?>a*))*/ (a PCRE_EXTRA facility).
3201 Version 1.00 18-Nov-97
3202 ----------------------
3204 1. Added compile-time macros to support systems such as SunOS4 which don't have
3205 memmove() or strerror() but have other things that can be used instead.
3207 2. Arranged that "make clean" removes the executables.
3210 Version 0.99 27-Oct-97
3211 ----------------------
3213 1. Fixed bug in code for optimizing classes with only one character. It was
3214 initializing a 32-byte map regardless, which could cause it to run off the end
3215 of the memory it had got.
3217 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3220 Version 0.98 22-Oct-97
3221 ----------------------
3223 1. Fixed bug in code for handling temporary memory usage when there are more
3224 back references than supplied space in the ovector. This could cause segfaults.
3227 Version 0.97 21-Oct-97
3228 ----------------------
3230 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3232 2. Optimized negated single characters not to use a bit map.
3234 3. Brought error texts together as macro definitions; clarified some of them;
3235 fixed one that was wrong - it said "range out of order" when it meant "invalid
3236 escape sequence".
3238 4. Changed some char * arguments to const char *.
3240 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3242 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3243 pcretest.
3246 Version 0.96 16-Oct-97
3247 ----------------------
3249 1. Added a simple "pgrep" utility to the distribution.
3251 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3252 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3253 where "ddd" means "one or more decimal digits".
3255 3. Fixed serious bug. If a pattern had a back reference, but the call to
3256 pcre_exec() didn't supply a large enough ovector to record the related
3257 identifying subpattern, the match always failed. PCRE now remembers the number
3258 of the largest back reference, and gets some temporary memory in which to save
3259 the offsets during matching if necessary, in order to ensure that
3260 backreferences always work.
3262 4. Increased the compatibility with Perl in a number of ways:
3264 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3265 to request this handling. The option can be set at compile or exec time.
3267 (b) $ matches before a terminating newline by default; an option
3268 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3269 mode). The option can be set at compile or exec time.
3271 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3272 the same as Perl's. If the decimal number it represents is less than 10
3273 or there aren't that many previous left capturing parentheses, an octal
3274 escape is read. Inside a character class, it's always an octal escape,
3275 even if it is a single digit.
3277 (d) An escaped but undefined alphabetic character is taken as a literal,
3278 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3279 escapes.
3281 (e) {0} is now permitted. (The previous item is removed from the compiled
3282 pattern).
3284 5. Changed all the names of code files so that the basic parts are no longer
3285 than 10 characters, and abolished the teeny "globals.c" file.
3287 6. Changed the handling of character classes; they are now done with a 32-byte
3288 bit map always.
3290 7. Added the -d and /D options to pcretest to make it possible to look at the
3291 internals of compilation without having to recompile pcre.
3294 Version 0.95 23-Sep-97
3295 ----------------------
3297 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3298 \x20 at the start of a run of normal characters. These were being treated as
3299 real characters, instead of the source characters being re-checked.
3302 Version 0.94 18-Sep-97
3303 ----------------------
3305 1. The functions are now thread-safe, with the caveat that the global variables
3306 containing pointers to malloc() and free() or alternative functions are the
3307 same for all threads.
3309 2. Get pcre_study() to generate a bitmap of initial characters for non-
3310 anchored patterns when this is possible, and use it if passed to pcre_exec().
3313 Version 0.93 15-Sep-97
3314 ----------------------
3316 1. /(b)|(:+)/ was computing an incorrect first character.
3318 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3319 but not actually doing anything yet.
3321 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3322 as Perl does (e.g. [-az] or [az-]).
3324 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3325 all possible positions.
3327 5. Split up into different modules to avoid including unneeded functions in a
3328 compiled binary. However, compile and exec are still in one module. The "study"
3329 function is split off.
3331 6. The character tables are now in a separate module whose source is generated
3332 by an auxiliary program - but can then be edited by hand if required. There are
3333 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3334 toupper() in the code.
3336 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3337 make them global. Abolish the function for setting them, as the caller can now
3338 set them directly.
3341 Version 0.92 11-Sep-97
3342 ----------------------
3344 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3345 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3347 2. Caseless matching was not working in character classes if the characters in
3348 the pattern were in upper case.
3350 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3352 4. Make PCRE_ANCHORED public and accept as a compile option.
3354 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3355 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3356 pass them.
3358 6. Give an error if bad option bits passed at compile or run time.
3360 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3361 pcretest to cause it to pass that flag.
3363 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3364 options, and the first character, if set.
3366 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3369 Version 0.91 10-Sep-97
3370 ----------------------
3372 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3373 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3375 2. PCRE was looping on encountering an indefinitely repeated back reference to
3376 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3377 Perl does - treats the match as successful.
3379 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5