ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 429 - (show annotations)
Tue Sep 1 16:10:16 2009 UTC (11 years, 11 months ago) by ph10
File size: 169779 byte(s)
Error occurred while calculating annotation data.
Add pcredemo man page, containing a listing of pcredemo.c.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.00 ??-???-??
5 ----------------------
7 1. The table for translating pcre_compile() error codes into POSIX error codes
8 was out-of-date, and there was no check on the pcre_compile() error code
9 being within the table. This could lead to an OK return being given in
10 error.
12 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
13 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
14 in a Windows environment.
16 3. The pcregrep --count option prints the count for each file even when it is
17 zero, as does GNU grep. However, pcregrep was also printing all files when
18 --files-with-matches was added. Now, when both options are given, it prints
19 counts only for those files that have at least one match. (GNU grep just
20 prints the file name in this circumstance, but including the count seems
21 more useful - otherwise, why use --count?) Also ensured that the
22 combination -clh just lists non-zero counts, with no names.
24 4. The long form of the pcregrep -F option was incorrectly implemented as
25 --fixed_strings instead of --fixed-strings. This is an incompatible change,
26 but it seems right to fix it, and I didn't think it was worth preserving
27 the old behaviour.
29 5. The command line items --regex=pattern and --regexp=pattern were not
30 recognized by pcregrep, which required --regex pattern or --regexp pattern
31 (with a space rather than an '='). The man page documented the '=' forms,
32 which are compatible with GNU grep; these now work.
34 6. No libpcreposix.pc file was created for pkg-config; there was just
35 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
37 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
38 when UCP support is not needed, by modifying the Python script that
39 generates it from Unicode data files. This should not matter if the module
40 is correctly used as a library, but I received one complaint about 50K of
41 unwanted data. My guess is that the person linked everything into his
42 program rather than using a library. Anyway, it does no harm.
44 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
45 was a minimum greater than 1 for a wide character in a possessive
46 repetition. Chaos could result.
48 9. The restrictions on what a pattern can contain when partial matching is
49 requested for pcre_exec() have been removed. All patterns can now be
50 partially matched by this function. In addition, if there are at least two
51 slots in the offset vector, the offsets of the first-encountered partial
52 match are set in them when PCRE_ERROR_PARTIAL is returned.
54 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
55 synonymous with PCRE_PARTIAL, for backwards compatibility, and
56 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
57 and may be more useful for multi-segment matching, especially with
58 pcre_exec().
60 11. Partial matching with pcre_exec() is now more intuitive. A partial match
61 used to be given if ever the end of the subject was reached; now it is
62 given only if matching could not proceed because another character was
63 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
64 string "Z", which now yields "no match" instead of "partial match". In the
65 case of pcre_dfa_exec(), "no match" is given if every matching path for the
66 final character ended with (*FAIL).
68 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
69 if the pattern had a "must contain" character that was already found in the
70 earlier partial match, unless partial matching was again requested. For
71 example, with the pattern /dog.(body)?/, the "must contain" character is
72 "g". If the first part-match was for the string "dog", restarting with
73 "sbody" failed.
75 13. Added a pcredemo man page, created automatically from the pcredemo.c file,
76 so that the demonstration program is easily available in environments where
77 PCRE has not been installed from source.
80 Version 7.9 11-Apr-09
81 ---------------------
83 1. When building with support for bzlib/zlib (pcregrep) and/or readline
84 (pcretest), all targets were linked against these libraries. This included
85 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
86 libraries. This caused unwanted dependencies to be created. This problem
87 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
88 pcretest is linked with readline.
90 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
91 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
92 moved outside it again, because FALSE and TRUE are already defined in AIX,
93 but BOOL is not.
95 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
96 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
98 4. The pcregrep documentation said spaces were inserted as well as colons (or
99 hyphens) following file names and line numbers when outputting matching
100 lines. This is not true; no spaces are inserted. I have also clarified the
101 wording for the --colour (or --color) option.
103 5. In pcregrep, when --colour was used with -o, the list of matching strings
104 was not coloured; this is different to GNU grep, so I have changed it to be
105 the same.
107 6. When --colo(u)r was used in pcregrep, only the first matching substring in
108 each matching line was coloured. Now it goes on to look for further matches
109 of any of the test patterns, which is the same behaviour as GNU grep.
111 7. A pattern that could match an empty string could cause pcregrep to loop; it
112 doesn't make sense to accept an empty string match in pcregrep, so I have
113 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
114 seems to be how GNU grep behaves.
116 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
117 start or after a newline", because the conditional assertion was not being
118 correctly handled. The rule now is that both the assertion and what follows
119 in the first alternative must satisfy the test.
121 9. If auto-callout was enabled in a pattern with a conditional group whose
122 condition was an assertion, PCRE could crash during matching, both with
123 pcre_exec() and pcre_dfa_exec().
125 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
126 used for matching.
128 11. Unicode property support in character classes was not working for
129 characters (bytes) greater than 127 when not in UTF-8 mode.
131 12. Added the -M command line option to pcretest.
133 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
135 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
137 16. Added comments and documentation about mis-use of no_arg in the C++
138 wrapper.
140 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
141 from Martin Jerabek that uses macro names for all relevant character and
142 string constants.
144 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
145 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
146 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
147 these, but not everybody uses configure.
149 19. A conditional group that had only one branch was not being correctly
150 recognized as an item that could match an empty string. This meant that an
151 enclosing group might also not be so recognized, causing infinite looping
152 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
153 with the subject "ab", where knowledge that the repeated group can match
154 nothing is needed in order to break the loop.
156 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
157 exec(), but without supplying a callout function, matching went wrong.
159 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
160 leak if the size of the offset vector was greater than 30. When the vector
161 is smaller, the saved offsets during recursion go onto a local stack
162 vector, but for larger vectors malloc() is used. It was failing to free
163 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
164 error, in fact).
166 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
167 heapframe that is used only when UTF-8 support is enabled. This caused no
168 problem, but was untidy.
170 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
171 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
172 included within another project.
174 24. Steven Van Ingelgem's patches to add more options to the CMake support,
175 slightly modified by me:
177 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
178 not building pcregrep.
180 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
181 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
183 25. Forward references, both numeric and by name, in patterns that made use of
184 duplicate group numbers, could behave incorrectly or give incorrect errors,
185 because when scanning forward to find the reference group, PCRE was not
186 taking into account the duplicate group numbers. A pattern such as
187 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
189 26. Changed a few more instances of "const unsigned char *" to USPTR, making
190 the feature of a custom pointer more persuasive (as requested by a user).
192 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
193 pcretest.c, inside #ifndefs, because it seems they are sometimes already
194 pre-defined.
196 28. Added support for (*UTF8) at the start of a pattern.
198 29. Arrange for flags added by the "release type" setting in CMake to be shown
199 in the configuration summary.
202 Version 7.8 05-Sep-08
203 ---------------------
205 1. Replaced UCP searching code with optimized version as implemented for Ad
206 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
207 stage table and inline lookup instead of a function, giving speed ups of 2
208 to 5 times on some simple patterns that I tested. Permission was given to
209 distribute the MultiStage2.py script that generates the tables (it's not in
210 the tarball, but is in the Subversion repository).
212 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
213 scripts.
215 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
216 a group with a zero qualifier. The result of the study could be incorrect,
217 or the function might crash, depending on the pattern.
219 4. Caseless matching was not working for non-ASCII characters in back
220 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
221 It now works when Unicode Property Support is available.
223 5. In pcretest, an escape such as \x{de} in the data was always generating
224 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
225 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
226 truncation.
228 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
230 7. Added two (int) casts to pcregrep when printing the difference of two
231 pointers, in case they are 64-bit values.
233 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
234 test 2 if it fails.
236 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
237 and a #define of that name to empty if it is not externally set. This is to
238 allow users of MSVC to set it if necessary.
240 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
241 the convenience functions in the pcre_get.c source file.
243 11. An option change at the start of a pattern that had top-level alternatives
244 could cause overwriting and/or a crash. This command provoked a crash in
245 some environments:
247 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
249 This potential security problem was recorded as CVE-2008-2371.
251 12. For a pattern where the match had to start at the beginning or immediately
252 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
253 pcre_dfa_exec() could read past the end of the passed subject if there was
254 no match. To help with detecting such bugs (e.g. with valgrind), I modified
255 pcretest so that it places the subject at the end of its malloc-ed buffer.
257 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
258 exec() might read past the end of the data buffer in UTF-8 mode.
260 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
261 the data contained the byte 0x85 as part of a UTF-8 character within its
262 first line. This applied both to normal and DFA matching.
264 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
265 /^[^d]*?$/8 failed to match "abc".
267 16. Added a missing copyright notice to pcrecpp_internal.h.
269 17. Make it more clear in the documentation that values returned from
270 pcre_exec() in ovector are byte offsets, not character counts.
272 18. Tidied a few places to stop certain compilers from issuing warnings.
274 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
275 supplied by Stefan Weber. I made a further small update for 7.8 because
276 there is a change of source arrangements: the pcre_searchfuncs.c module is
277 replaced by pcre_ucd.c.
280 Version 7.7 07-May-08
281 ---------------------
283 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
284 a string to a long long, pretend we don't even have a long long." This is
285 done by checking for the strtoq, strtoll, and _strtoi64 functions.
287 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
288 pre-7.6 versions, which defined a global no_arg variable instead of putting
289 it in the RE class. (See also #8 below.)
291 3. Remove a line of dead code, identified by coverity and reported by Nuno
292 Lopes.
294 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
296 (1) The include/exclude patterns were being applied to the whole pathnames
297 of files, instead of just to the final components.
299 (2) If there was more than one level of directory, the subdirectories were
300 skipped unless they satisfied the include/exclude conditions. This is
301 inconsistent with GNU grep (and could even be seen as contrary to the
302 pcregrep specification - which I improved to make it absolutely clear).
303 The action now is always to scan all levels of directory, and just
304 apply the include/exclude patterns to regular files.
306 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
307 --exclude_dir in the tests to avoid scanning .svn directories.
309 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
310 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
311 doesn't support NULs in patterns.
313 7. Added some missing "const"s to declarations of static tables in
314 pcre_compile.c and pcre_dfa_exec.c.
316 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
317 caused by fix #2 above. (Subsequently also a second patch to fix the
318 first patch. And a third patch - this was a messy problem.)
320 9. Applied Craig's patch to remove the use of push_back().
322 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
323 matching function regexec().
325 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
326 which, however, unlike Perl's \g{...}, are subroutine calls, not back
327 references. PCRE supports relative numbers with this syntax (I don't think
328 Oniguruma does).
330 12. Previously, a group with a zero repeat such as (...){0} was completely
331 omitted from the compiled regex. However, this means that if the group
332 was called as a subroutine from elsewhere in the pattern, things went wrong
333 (an internal error was given). Such groups are now left in the compiled
334 pattern, with a new opcode that causes them to be skipped at execution
335 time.
337 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
338 to the way PCRE behaves:
340 (a) A lone ] character is dis-allowed (Perl treats it as data).
342 (b) A back reference to an unmatched subpattern matches an empty string
343 (Perl fails the current match path).
345 (c) A data ] in a character class must be notated as \] because if the
346 first data character in a class is ], it defines an empty class. (In
347 Perl it is not possible to have an empty class.) The empty class []
348 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
349 The negative empty class [^] matches any one character, independently
350 of the DOTALL setting.
352 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
353 non-existent subpattern following a character class starting with ']' and
354 containing () gave an internal compiling error instead of "reference to
355 non-existent subpattern". Fortunately, when the pattern did exist, the
356 compiled code was correct. (When scanning forwards to check for the
357 existencd of the subpattern, it was treating the data ']' as terminating
358 the class, so got the count wrong. When actually compiling, the reference
359 was subsequently set up correctly.)
361 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
362 it was being rejected as not supported by pcre_dfa_exec(), even though
363 other assertions are supported. I have made pcre_dfa_exec() support
364 (*FAIL).
366 16. The implementation of 13c above involved the invention of a new opcode,
367 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
368 cannot be changed at match time, I realized I could make a small
369 improvement to matching performance by compiling OP_ALLANY instead of
370 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
371 on the OP_ANY path.
373 17. Compiling pcretest on Windows with readline support failed without the
374 following two fixes: (1) Make the unistd.h include conditional on
375 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
377 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
378 ncurses library to be included for pcretest when ReadLine support is
379 requested, but also to allow for it to be overridden. This patch came from
380 Daniel Bergström.
382 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
383 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
384 any errors with the current Unicode tables. Thanks to Peter Kankowski for
385 spotting this.
388 Version 7.6 28-Jan-08
389 ---------------------
391 1. A character class containing a very large number of characters with
392 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
393 overflow.
395 2. Patch to cut out the "long long" test in pcrecpp_unittest when
396 HAVE_LONG_LONG is not defined.
398 3. Applied Christian Ehrlicher's patch to update the CMake build files to
399 bring them up to date and include new features. This patch includes:
401 - Fixed PH's badly added libz and libbz2 support.
402 - Fixed a problem with static linking.
403 - Added pcredemo. [But later removed - see 7 below.]
404 - Fixed dftables problem and added an option.
405 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
407 - Added readline support for pcretest.
408 - Added an listing of the option settings after cmake has run.
410 4. A user submitted a patch to Makefile that makes it easy to create
411 "pcre.dll" under mingw when using Configure/Make. I added stuff to
412 Makefile.am that cause it to include this special target, without
413 affecting anything else. Note that the same mingw target plus all
414 the other distribution libraries and programs are now supported
415 when configuring with CMake (see 6 below) instead of with
416 Configure/Make.
418 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
419 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
420 exported in the Windows port". It has not yet been confirmed that the patch
421 solves the problem, but it does no harm.
423 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
424 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
425 with CMake, and also correct the comment about stack recursion.
427 7. Remove the automatic building of pcredemo from the ./configure system and
428 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
429 of a program that users should build themselves after PCRE is installed, so
430 building it automatically is not really right. What is more, it gave
431 trouble in some build environments.
433 8. Further tidies to CMakeLists.txt from Sheri and Christian.
436 Version 7.5 10-Jan-08
437 ---------------------
439 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
440 values in parens when parsing an RE using the C++ wrapper."
442 2. Negative specials like \S did not work in character classes in UTF-8 mode.
443 Characters greater than 255 were excluded from the class instead of being
444 included.
446 3. The same bug as (2) above applied to negated POSIX classes such as
447 [:^space:].
449 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
450 defined or documented. It seems to have been a typo for PCRE_STATIC, so
451 I have changed it.
453 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
454 first named subpattern) and a construct such as (?&a) would reference the
455 first named subpattern whose name started with "a" (in other words, the
456 length check was missing). Both these problems are fixed. "Subpattern name
457 expected" is now given for (?&) (a zero-length name), and this patch also
458 makes it give the same error for \k'' (previously it complained that that
459 was a reference to a non-existent subpattern).
461 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
462 this is right because (?- can be followed by option settings as well as by
463 digits. I have, however, made the messages clearer.
465 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
466 than the number used in the conditional) now cause a compile-time error.
467 This is actually not compatible with Perl, which accepts such patterns, but
468 treats the conditional as always being FALSE (as PCRE used to), but it
469 seems to me that giving a diagnostic is better.
471 8. Change "alphameric" to the more common word "alphanumeric" in comments
472 and messages.
474 9. Fix two occurrences of "backslash" in comments that should have been
475 "backspace".
477 10. Remove two redundant lines of code that can never be obeyed (their function
478 was moved elsewhere).
480 11. The program that makes PCRE's Unicode character property table had a bug
481 which caused it to generate incorrect table entries for sequences of
482 characters that have the same character type, but are in different scripts.
483 It amalgamated them into a single range, with the script of the first of
484 them. In other words, some characters were in the wrong script. There were
485 thirteen such cases, affecting characters in the following ranges:
487 U+002b0 - U+002c1
488 U+0060c - U+0060d
489 U+0061e - U+00612
490 U+0064b - U+0065e
491 U+0074d - U+0076d
492 U+01800 - U+01805
493 U+01d00 - U+01d77
494 U+01d9b - U+01dbf
495 U+0200b - U+0200f
496 U+030fc - U+030fe
497 U+03260 - U+0327f
498 U+0fb46 - U+0fbb1
499 U+10450 - U+1049d
501 12. The -o option (show only the matching part of a line) for pcregrep was not
502 compatible with GNU grep in that, if there was more than one match in a
503 line, it showed only the first of them. It now behaves in the same way as
504 GNU grep.
506 13. If the -o and -v options were combined for pcregrep, it printed a blank
507 line for every non-matching line. GNU grep prints nothing, and pcregrep now
508 does the same. The return code can be used to tell if there were any
509 non-matching lines.
511 14. Added --file-offsets and --line-offsets to pcregrep.
513 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
514 infinitely looping recursion. The bug was that positive lookaheads were not
515 being skipped when checking for a possible empty match (negative lookaheads
516 and both kinds of lookbehind were skipped).
518 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
519 inclusion of <windows.h> to before rather than after the definition of
520 INVALID_FILE_ATTRIBUTES (patch from David Byron).
522 17. Specifying a possessive quantifier with a specific limit for a Unicode
523 character property caused pcre_compile() to compile bad code, which led at
524 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
525 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
526 caused the error; without that there was no problem.
528 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
530 19. Added --enable-pcretest-libreadline.
532 20. In pcrecpp.cc, the variable 'count' was incremented twice in
533 RE::GlobalReplace(). As a result, the number of replacements returned was
534 double what it should be. I removed one of the increments, but Craig sent a
535 later patch that removed the other one (the right fix) and added unit tests
536 that check the return values (which was not done before).
538 21. Several CMake things:
540 (1) Arranged that, when cmake is used on Unix, the libraries end up with
541 the names libpcre and libpcreposix, not just pcre and pcreposix.
543 (2) The above change means that pcretest and pcregrep are now correctly
544 linked with the newly-built libraries, not previously installed ones.
548 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
549 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
550 UTF-8 newline character). The key issue is that the pattern starts .*;
551 this means that the match must be either at the beginning, or after a
552 newline. The bug was in the code for advancing after a failed match and
553 checking that the new position followed a newline. It was not taking
554 account of UTF-8 characters correctly.
556 23. PCRE was behaving differently from Perl in the way it recognized POSIX
557 character classes. PCRE was not treating the sequence [:...:] as a
558 character class unless the ... were all letters. Perl, however, seems to
559 allow any characters between [: and :], though of course it rejects as
560 unknown any "names" that contain non-letters, because all the known class
561 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
562 for example, whereas PCRE did not - it did not recognize a POSIX character
563 class. This seemed a bit dangerous, so the code has been changed to be
564 closer to Perl. The behaviour is not identical to Perl, because PCRE will
565 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
566 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
567 Perl does, and where it didn't before.
569 24. Rewrite so as to remove the single use of %n from pcregrep because in some
570 Windows environments %n is disabled by default.
573 Version 7.4 21-Sep-07
574 ---------------------
576 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
577 means that a class such as [\s] counted as "explicit reference to CR or
578 LF". That isn't really right - the whole point of the change was to try to
579 help when there was an actual mention of one of the two characters. So now
580 the change happens only if \r or \n (or a literal CR or LF) character is
581 encountered.
583 2. The 32-bit options word was also used for 6 internal flags, but the numbers
584 of both had grown to the point where there were only 3 bits left.
585 Fortunately, there was spare space in the data structure, and so I have
586 moved the internal flags into a new 16-bit field to free up more option
587 bits.
589 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
590 but did not set the internal JCHANGED flag - either of these is enough to
591 control the way the "get" function works - but the PCRE_INFO_JCHANGED
592 facility is supposed to tell if (?J) was ever used, so now (?J) at the
593 start sets both bits.
595 4. Added options (at build time, compile time, exec time) to change \R from
596 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
598 5. doc/pcresyntax.html was missing from the distribution.
600 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
601 compatibility, even though it is no longer used.
603 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
604 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
605 windows.h file is present (where different names are used). [This was
606 reversed later after testing - see 16 below.]
608 8. Changed all #include <config.h> to #include "config.h". There were also
609 some further <pcre.h> cases that I changed to "pcre.h".
611 9. When pcregrep was used with the --colour option, it missed the line ending
612 sequence off the lines that it output.
614 10. It was pointed out to me that arrays of string pointers cause lots of
615 relocations when a shared library is dynamically loaded. A technique of
616 using a single long string with a table of offsets can drastically reduce
617 these. I have refactored PCRE in four places to do this. The result is
618 dramatic:
620 Originally: 290
621 After changing UCP table: 187
622 After changing error message table: 43
623 After changing table of "verbs" 36
624 After changing table of Posix names 22
626 Thanks to the folks working on Gregex for glib for this insight.
628 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
629 unicode-properties was also set.
631 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
633 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
634 checked only for CRLF.
636 14. Added casts to pcretest.c to avoid compiler warnings.
638 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
640 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
641 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
642 entirely. This removes changes made in 7 above.
644 17. The CMake files have been updated, and there is now more information about
645 building with CMake in the NON-UNIX-USE document.
648 Version 7.3 28-Aug-07
649 ---------------------
651 1. In the rejigging of the build system that eventually resulted in 7.1, the
652 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
653 brackets there is not right, since it causes compilers to look for an
654 installed pcre.h, not the version that is in the source that is being
655 compiled (which of course may be different). I have changed it back to:
657 #include "pcre.h"
659 I have a vague recollection that the change was concerned with compiling in
660 different directories, but in the new build system, that is taken care of
661 by the VPATH setting the Makefile.
663 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
664 when the subject happened to end in the byte 0x85 (e.g. if the last
665 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
666 characters but of course it shouldn't be taken as a newline when it is part
667 of another character. The bug was that, for an unlimited repeat of . in
668 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
669 characters when looking for a newline.
671 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
673 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
674 in debug output.
676 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
677 long printing in the pcrecpp unittest when running under MinGW.
679 6. ESC_K was left out of the EBCDIC table.
681 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
682 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
683 limit also applies to "virtual nesting" when a pattern is recursive, and in
684 this case 1000 isn't so big. I have been able to remove this limit at the
685 expense of backing off one optimization in certain circumstances. Normally,
686 when pcre_exec() would call its internal match() function recursively and
687 immediately return the result unconditionally, it uses a "tail recursion"
688 feature to save stack. However, when a subpattern that can match an empty
689 string has an unlimited repetition quantifier, it no longer makes this
690 optimization. That gives it a stack frame in which to save the data for
691 checking that an empty string has been matched. Previously this was taken
692 from the 1000-entry workspace that had been reserved. So now there is no
693 explicit limit, but more stack is used.
695 8. Applied Daniel's patches to solve problems with the import/export magic
696 syntax that is required for Windows, and which was going wrong for the
697 pcreposix and pcrecpp parts of the library. These were overlooked when this
698 problem was solved for the main library.
700 9. There were some crude static tests to avoid integer overflow when computing
701 the size of patterns that contain repeated groups with explicit upper
702 limits. As the maximum quantifier is 65535, the maximum group length was
703 set at 30,000 so that the product of these two numbers did not overflow a
704 32-bit integer. However, it turns out that people want to use groups that
705 are longer than 30,000 bytes (though not repeat them that many times).
706 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
707 made it possible to implement the integer overflow checks in a much more
708 dynamic way, which I have now done. The artificial limitation on group
709 length has been removed - we now have only the limit on the total length of
710 the compiled pattern, which depends on the LINK_SIZE setting.
712 10. Fixed a bug in the documentation for get/copy named substring when
713 duplicate names are permitted. If none of the named substrings are set, the
714 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
715 empty string.
717 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
718 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
719 because the ] is interpreted as the first data character and the
720 terminating ] is not found. PCRE has been made compatible with Perl in this
721 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
722 cause memory overwriting.
724 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
725 string has been matched (to stop an infinite loop). It was not recognizing
726 a conditional subpattern that could match an empty string if that
727 subpattern was within another subpattern. For example, it looped when
728 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
729 condition was not nested. This bug has been fixed.
731 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
732 past the start of the subject in the presence of bytes with the top bit
733 set, for example "\x8aBCD".
735 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
736 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
738 14. Optimized (?!) to (*FAIL).
740 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
741 This restricts code points to be within the range 0 to 0x10FFFF, excluding
742 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
743 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
744 does: it's just the validity check that is more restrictive.
746 16. Inserted checks for integer overflows during escape sequence (backslash)
747 processing, and also fixed erroneous offset values for syntax errors during
748 backslash processing.
750 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
751 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
753 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
754 caused an overrun.
756 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
757 something other than just ASCII characters) inside a group that had an
758 unlimited repeat caused a loop at compile time (while checking to see
759 whether the group could match an empty string).
761 20. Debugging a pattern containing \p or \P could cause a crash. For example,
762 [\P{Any}] did so. (Error in the code for printing property names.)
764 21. An orphan \E inside a character class could cause a crash.
766 22. A repeated capturing bracket such as (A)? could cause a wild memory
767 reference during compilation.
769 23. There are several functions in pcre_compile() that scan along a compiled
770 expression for various reasons (e.g. to see if it's fixed length for look
771 behind). There were bugs in these functions when a repeated \p or \P was
772 present in the pattern. These operators have additional parameters compared
773 with \d, etc, and these were not being taken into account when moving along
774 the compiled data. Specifically:
776 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
777 length.
779 (b) An item such as \pL+ within a repeated group could cause crashes or
780 loops.
782 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
783 "reference to non-existent subpattern" error.
785 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
787 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
788 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
790 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
792 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
793 character were causing crashes (broken optimization).
795 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
796 \p or \P) caused a compile-time loop.
798 28. More problems have arisen in unanchored patterns when CRLF is a valid line
799 break. For example, the unstudied pattern [\r\n]A does not match the string
800 "\r\nA" because change 7.0/46 below moves the current point on by two
801 characters after failing to match at the start. However, the pattern \nA
802 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
803 the same is true. There doesn't seem any very clean way out of this, but
804 what I have chosen to do makes the common cases work: PCRE now takes note
805 of whether there can be an explicit match for \r or \n anywhere in the
806 pattern, and if so, 7.0/46 no longer applies. As part of this change,
807 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
808 pattern has explicit CR or LF references.
810 29. Added (*CR) etc for changing newline setting at start of pattern.
813 Version 7.2 19-Jun-07
814 ---------------------
816 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
817 which is apparently normally available under Windows.
819 2. Re-jig the pcregrep tests with different newline settings in an attempt
820 to make them independent of the local environment's newline setting.
822 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
824 4. Some of the "internals" tests were previously cut out when the link size
825 was not 2, because the output contained actual offsets. The recent new
826 "Z" feature of pcretest means that these can be cut out, making the tests
827 usable with all link sizes.
829 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
830 stack recursion. This gives a massive performance boost under BSD, but just
831 a small improvement under Linux. However, it saves one field in the frame
832 in all cases.
834 6. Added more features from the forthcoming Perl 5.10:
836 (a) (?-n) (where n is a string of digits) is a relative subroutine or
837 recursion call. It refers to the nth most recently opened parentheses.
839 (b) (?+n) is also a relative subroutine call; it refers to the nth next
840 to be opened parentheses.
842 (c) Conditions that refer to capturing parentheses can be specified
843 relatively, for example, (?(-2)... or (?(+3)...
845 (d) \K resets the start of the current match so that everything before
846 is not part of it.
848 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
850 (f) \g{name} is another synonym - part of Perl 5.10's unification of
851 reference syntax.
853 (g) (?| introduces a group in which the numbering of parentheses in each
854 alternative starts with the same number.
856 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
858 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
861 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
862 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
863 for detecting groups that can match an empty string.
865 9. A pattern with a very large number of alternatives (more than several
866 hundred) was running out of internal workspace during the pre-compile
867 phase, where pcre_compile() figures out how much memory will be needed. A
868 bit of new cunning has reduced the workspace needed for groups with
869 alternatives. The 1000-alternative test pattern now uses 12 bytes of
870 workspace instead of running out of the 4096 that are available.
872 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
874 11. Applied patch from Google to remove an optimization that didn't quite work.
875 The report of the bug said:
877 pcrecpp::RE("a*").FullMatch("aaa") matches, while
878 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
879 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
881 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
882 it matched the wrong number of bytes.
885 Version 7.1 24-Apr-07
886 ---------------------
888 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
889 that is more "standard", making use of automake and other Autotools. There
890 is some re-arrangement of the files and adjustment of comments consequent
891 on this.
893 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
894 for recursive directory scanning broke on some systems because the files
895 are not scanned in any specific order and on different systems the order
896 was different. A call to "sort" has been inserted into RunGrepTest for the
897 approprate test as a short-term fix. In the longer term there may be an
898 alternative.
900 3. I had an email from Eric Raymond about problems translating some of PCRE's
901 man pages to HTML (despite the fact that I distribute HTML pages, some
902 people do their own conversions for various reasons). The problems
903 concerned the use of low-level troff macros .br and .in. I have therefore
904 removed all such uses from the man pages (some were redundant, some could
905 be replaced by .nf/.fi pairs). The 132html script that I use to generate
906 HTML has been updated to handle .nf/.fi and to complain if it encounters
907 .br or .in.
909 4. Updated comments in configure.ac that get placed in config.h.in and also
910 arranged for config.h to be included in the distribution, with the name
911 config.h.generic, for the benefit of those who have to compile without
912 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
914 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
915 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
916 makevp.bat for the current PCRE, using the additional files
917 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
919 6. A Windows user reported a minor discrepancy with test 2, which turned out
920 to be caused by a trailing space on an input line that had got lost in his
921 copy. The trailing space was an accident, so I've just removed it.
923 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
924 that is needed.
926 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
927 as "const" (a) because they are and (b) because it helps the PHP
928 maintainers who have recently made a script to detect big data structures
929 in the php code that should be moved to the .rodata section. I remembered
930 to update Builducptable as well, so it won't revert if ucptable.h is ever
931 re-created.
933 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
934 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
935 order to be able to cut out the UTF-8 tables in the latter when UTF-8
936 support is not required. This saves 1.5-2K of code, which is important in
937 some applications.
939 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
940 so as not to refer to the tables, even though these functions will never be
941 called when UTF-8 support is disabled. Otherwise there are problems with a
942 shared library.
944 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
946 (a) It was defining its arguments as char * instead of void *.
948 (b) It was assuming that all moves were upwards in memory; this was true
949 a long time ago when I wrote it, but is no longer the case.
951 The emulated memove() is provided for those environments that have neither
952 memmove() nor bcopy(). I didn't think anyone used it these days, but that
953 is clearly not the case, as these two bugs were recently reported.
955 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
956 and Detrail to create the HTML documentation, the .txt form of the man
957 pages, and it removes trailing spaces from listed files. It also creates
958 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
959 case, it wraps all the #defines with #ifndefs. This script should be run
960 before "make dist".
962 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
963 with Unicode property support.
965 (a) For a maximizing quantifier, if the two different cases of the
966 character were of different lengths in their UTF-8 codings (there are
967 some cases like this - I found 11), and the matching function had to
968 back up over a mixture of the two cases, it incorrectly assumed they
969 were both the same length.
971 (b) When PCRE was configured to use the heap rather than the stack for
972 recursion during matching, it was not correctly preserving the data for
973 the other case of a UTF-8 character when checking ahead for a match
974 while processing a minimizing repeat. If the check also involved
975 matching a wide character, but failed, corruption could cause an
976 erroneous result when trying to check for a repeat of the original
977 character.
979 13. Some tidying changes to the testing mechanism:
981 (a) The RunTest script now detects the internal link size and whether there
982 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
983 values substituted by "configure". (The RunGrepTest script already did
984 this for UTF-8.) The configure.ac script no longer substitutes the
985 relevant variables.
987 (b) The debugging options /B and /D in pcretest show the compiled bytecode
988 with length and offset values. This means that the output is different
989 for different internal link sizes. Test 2 is skipped for link sizes
990 other than 2 because of this, bypassing the problem. Unfortunately,
991 there was also a test in test 3 (the locale tests) that used /B and
992 failed for link sizes other than 2. Rather than cut the whole test out,
993 I have added a new /Z option to pcretest that replaces the length and
994 offset values with spaces. This is now used to make test 3 independent
995 of link size. (Test 2 will be tidied up later.)
997 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
998 segmentation fault instead of returning the appropriate error message.
1000 15. In multiline mode when the newline sequence was set to "any", the pattern
1001 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1002 This doesn't seem right; it now treats the CRLF combination as the line
1003 ending, and so does not match in that case. It's only a pattern such as ^$
1004 that would hit this one: something like ^ABC$ would have failed after \r
1005 and then tried again after \r\n.
1007 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1008 in an attempt to make files that differ only in their line terminators
1009 compare equal. This works on Linux.
1011 17. Under certain error circumstances pcregrep might try to free random memory
1012 as it exited. This is now fixed, thanks to valgrind.
1014 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1015 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1016 was because its rules for how to advance for /g after matching an empty
1017 string at the end of a line did not allow for this case. They now check for
1018 it specially.
1020 20. pcretest is supposed to handle patterns and data of any length, by
1021 extending its buffers when necessary. It was getting this wrong when the
1022 buffer for a data line had to be extended.
1024 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1025 CRLF as a newline sequence.
1027 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1028 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1029 I have nevertheless tidied it up.
1031 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1033 24. Added a man page for pcre-config.
1036 Version 7.0 19-Dec-06
1037 ---------------------
1039 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1040 moving to gcc 4.1.1.
1042 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1043 sys/time.h, which is documented as needed for this function. It doesn't
1044 seem to matter on Linux, but it showed up on some releases of OS X.
1046 3. It seems that there are systems where bytes whose values are greater than
1047 127 match isprint() in the "C" locale. The "C" locale should be the
1048 default when a C program starts up. In most systems, only ASCII printing
1049 characters match isprint(). This difference caused the output from pcretest
1050 to vary, making some of the tests fail. I have changed pcretest so that:
1052 (a) When it is outputting text in the compiled version of a pattern, bytes
1053 other than 32-126 are always shown as hex escapes.
1055 (b) When it is outputting text that is a matched part of a subject string,
1056 it does the same, unless a different locale has been set for the match
1057 (using the /L modifier). In this case, it uses isprint() to decide.
1059 4. Fixed a major bug that caused incorrect computation of the amount of memory
1060 required for a compiled pattern when options that changed within the
1061 pattern affected the logic of the preliminary scan that determines the
1062 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1063 that the computed length was too small. The symptoms of this bug were
1064 either the PCRE error "internal error: code overflow" from pcre_compile(),
1065 or a glibc crash with a message such as "pcretest: free(): invalid next
1066 size (fast)". Examples of patterns that provoked this bug (shown in
1067 pcretest format) are:
1069 /(?-x: )/x
1070 /(?x)(?-x: \s*#\s*)/
1071 /((?i)[\x{c0}])/8
1072 /(?i:[\x{c0}])/8
1074 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1075 is now done differently.
1077 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1078 wrapper classes; (b) implement a new function in the C++ scanner that is
1079 more efficient than the old way of doing things because it avoids levels of
1080 recursion in the regex matching; (c) add a paragraph to the documentation
1081 for the FullMatch() function.
1083 6. The escape sequence \n was being treated as whatever was defined as
1084 "newline". Not only was this contrary to the documentation, which states
1085 that \n is character 10 (hex 0A), but it also went horribly wrong when
1086 "newline" was defined as CRLF. This has been fixed.
1088 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1089 was being set to -1 for the "end of line" case (supposedly a value that no
1090 character can have). Though this value is never used (the check for end of
1091 line is "zero bytes in current character"), it caused compiler complaints.
1092 I've changed it to 0xffffffff.
1094 8. In pcre_version.c, the version string was being built by a sequence of
1095 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1096 string (as it is for production releases) called a macro with an empty
1097 argument. The C standard says the result of this is undefined. The gcc
1098 compiler treats it as an empty string (which was what was wanted) but it is
1099 reported that Visual C gives an error. The source has been hacked around to
1100 avoid this problem.
1102 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1103 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1104 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1105 of them did).
1107 10. Originally, pcretest opened its input and output without "b"; then I was
1108 told that "b" was needed in some environments, so it was added for release
1109 5.0 to both the input and output. (It makes no difference on Unix-like
1110 systems.) Later I was told that it is wrong for the input on Windows. I've
1111 now abstracted the modes into two macros, to make it easier to fiddle with
1112 them, and removed "b" from the input mode under Windows.
1114 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1116 12. Added -help and --help to pcretest as an official way of being reminded
1117 of the options.
1119 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1120 and pcrecpp.cc because they annoy compilers at high warning levels.
1122 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1124 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1125 scripts are not C programs :-) and which was not noticed because it works
1126 on Linux.
1128 16. pcretest is supposed to handle any length of pattern and data line (as one
1129 line or as a continued sequence of lines) by extending its input buffer if
1130 necessary. This feature was broken for very long pattern lines, leading to
1131 a string of junk being passed to pcre_compile() if the pattern was longer
1132 than about 50K.
1134 17. I have done a major re-factoring of the way pcre_compile() computes the
1135 amount of memory needed for a compiled pattern. Previously, there was code
1136 that made a preliminary scan of the pattern in order to do this. That was
1137 OK when PCRE was new, but as the facilities have expanded, it has become
1138 harder and harder to keep it in step with the real compile phase, and there
1139 have been a number of bugs (see for example, 4 above). I have now found a
1140 cunning way of running the real compile function in a "fake" mode that
1141 enables it to compute how much memory it would need, while actually only
1142 ever using a few hundred bytes of working memory and without too many
1143 tests of the mode. This should make future maintenance and development
1144 easier. A side effect of this work is that the limit of 200 on the nesting
1145 depth of parentheses has been removed (though this was never a serious
1146 limitation, I suspect). However, there is a downside: pcre_compile() now
1147 runs more slowly than before (30% or more, depending on the pattern). I
1148 hope this isn't a big issue. There is no effect on runtime performance.
1150 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1151 newline (only possible for the last line of a file) and it was a
1152 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1154 19. Added additional timing features to pcretest. (1) The -tm option now times
1155 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1156 separate command line item, by a number that specifies the number of
1157 repeats to use when timing. The default is 50000; this gives better
1158 precision, but takes uncomfortably long for very large patterns.
1160 20. Extended pcre_study() to be more clever in cases where a branch of a
1161 subpattern has no definite first character. For example, (a*|b*)[cd] would
1162 previously give no result from pcre_study(). Now it recognizes that the
1163 first character must be a, b, c, or d.
1165 21. There was an incorrect error "recursive call could loop indefinitely" if
1166 a subpattern (or the entire pattern) that was being tested for matching an
1167 empty string contained only one non-empty item after a nested subpattern.
1168 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1169 incorrectly, because the \d was being skipped in the check.
1171 22. The pcretest program now has a new pattern option /B and a command line
1172 option -b, which is equivalent to adding /B to every pattern. This causes
1173 it to show the compiled bytecode, without the additional information that
1174 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1175 is the same as /B/I).
1177 23. A new optimization is now able automatically to treat some sequences such
1178 as a*b as a*+b. More specifically, if something simple (such as a character
1179 or a simple class like \d) has an unlimited quantifier, and is followed by
1180 something that cannot possibly match the quantified thing, the quantifier
1181 is automatically "possessified".
1183 24. A recursive reference to a subpattern whose number was greater than 39
1184 went wrong under certain circumstances in UTF-8 mode. This bug could also
1185 have affected the operation of pcre_study().
1187 25. Realized that a little bit of performance could be had by replacing
1188 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1190 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1192 27. Possessive quantifiers such as a++ were previously implemented by turning
1193 them into atomic groups such as ($>a+). Now they have their own opcodes,
1194 which improves performance. This includes the automatically created ones
1195 from 23 above.
1197 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1198 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1199 the first matched character to be a colon. This applied both to named and
1200 numbered groups.
1202 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1204 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1205 building PCRE on Windows easier, so I have included it in the distribution.
1207 31. There is now a check in pcretest against a ridiculously large number being
1208 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1209 loop, the loop is abandoned.
1211 32. Forward references to subpatterns in conditions such as (?(2)...) where
1212 subpattern 2 is defined later cause pcre_compile() to search forwards in
1213 the pattern for the relevant set of parentheses. This search went wrong
1214 when there were unescaped parentheses in a character class, parentheses
1215 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1217 33. "Subroutine" calls and backreferences were previously restricted to
1218 referencing subpatterns earlier in the regex. This restriction has now
1219 been removed.
1221 34. Added a number of extra features that are going to be in Perl 5.10. On the
1222 whole, these are just syntactic alternatives for features that PCRE had
1223 previously implemented using the Python syntax or my own invention. The
1224 other formats are all retained for compatibility.
1226 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1227 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1228 also .NET compatible.
1230 (b) A recursion or subroutine call to a named group can now be defined as
1231 (?&name) as well as (?P>name).
1233 (c) A backreference to a named group can now be defined as \k<name> or
1234 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1235 5.10, are also .NET compatible.
1237 (d) A conditional reference to a named group can now use the syntax
1238 (?(<name>) or (?('name') as well as (?(name).
1240 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1241 groups (named and numbered) that are never evaluated inline, but can be
1242 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1243 is always false. There may be only one alternative in such a group.
1245 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1246 as the simple (?(R). The condition is true only if the most recent
1247 recursion is that of the given number or name. It does not search out
1248 through the entire recursion stack.
1250 (g) The escape \gN or \g{N} has been added, where N is a positive or
1251 negative number, specifying an absolute or relative reference.
1253 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1254 some "unreachable code" warnings.
1256 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1257 things, this adds five new scripts.
1259 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1260 There were also incompatibilities regarding the handling of \Q..\E inside
1261 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1262 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1264 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1265 matches an empty string, and forcibly breaks the loop. There were bugs in
1266 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1267 against aaaa the result was just "a" rather than "aaaa", for example. Two
1268 separate and independent bugs (that affected different cases) have been
1269 fixed.
1271 39. Refactored the code to abolish the use of different opcodes for small
1272 capturing bracket numbers. This is a tidy that I avoided doing when I
1273 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1274 The new approach is not only tidier, it makes it possible to reduce the
1275 memory needed to fix the previous bug (38).
1277 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1278 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1279 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1280 mode.
1282 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1283 report.
1285 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1286 copy construction and assignment in the C++ wrapper.
1288 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1289 couple of bugs that could have given wrong results in the "--newline=crlf"
1290 case.
1292 44. Added a number of casts and did some reorganization of signed/unsigned int
1293 variables following suggestions from Dair Grant. Also renamed the variable
1294 "this" as "item" because it is a C++ keyword.
1296 45. Arranged for dftables to add
1298 #include "pcre_internal.h"
1300 to pcre_chartables.c because without it, gcc 4.x may remove the array
1301 definition from the final binary if PCRE is built into a static library and
1302 dead code stripping is activated.
1304 46. For an unanchored pattern, if a match attempt fails at the start of a
1305 newline sequence, and the newline setting is CRLF or ANY, and the next two
1306 characters are CRLF, advance by two characters instead of one.
1309 Version 6.7 04-Jul-06
1310 ---------------------
1312 1. In order to handle tests when input lines are enormously long, pcretest has
1313 been re-factored so that it automatically extends its buffers when
1314 necessary. The code is crude, but this _is_ just a test program. The
1315 default size has been increased from 32K to 50K.
1317 2. The code in pcre_study() was using the value of the re argument before
1318 testing it for NULL. (Of course, in any sensible call of the function, it
1319 won't be NULL.)
1321 3. The memmove() emulation function in pcre_internal.h, which is used on
1322 systems that lack both memmove() and bcopy() - that is, hardly ever -
1323 was missing a "static" storage class specifier.
1325 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1326 containing an extended class (one that cannot be represented by a bitmap
1327 because it contains high-valued characters or Unicode property items, e.g.
1328 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1329 pattern, but PCRE should not loop if you do not (it no longer does).
1330 [Detail: two cases were found: (a) a repeated subpattern containing an
1331 extended class; (b) a recursive reference to a subpattern that followed a
1332 previous extended class. It wasn't skipping over the extended class
1333 correctly when UTF-8 mode was not set.]
1335 5. A negated single-character class was not being recognized as fixed-length
1336 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1337 compile error "lookbehind assertion is not fixed length".
1339 6. The RunPerlTest auxiliary script was showing an unexpected difference
1340 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1341 write a Perl script that can interpret lines of an input file either as
1342 byte characters or as UTF-8, which is what "perltest" was being required to
1343 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1344 can't do is switch easily at run time between having the "use utf8;" pragma
1345 or not. In the end, I fudged it by using the RunPerlTest script to insert
1346 "use utf8;" explicitly for the UTF-8 tests.
1348 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1349 the end of the subject string, contrary to the documentation and to what
1350 Perl does. This was true of both matching functions. Now it matches only at
1351 the start of the subject and immediately after *internal* newlines.
1353 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1354 a pointer to an int instead of a pointer to an unsigned long int. This
1355 caused problems on 64-bit systems.
1357 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1358 instance of the 'standard' template library not being so standard".
1360 10. There was no check on the number of named subpatterns nor the maximum
1361 length of a subpattern name. The product of these values is used to compute
1362 the size of the memory block for a compiled pattern. By supplying a very
1363 long subpattern name and a large number of named subpatterns, the size
1364 computation could be caused to overflow. This is now prevented by limiting
1365 the length of names to 32 characters, and the number of named subpatterns
1366 to 10,000.
1368 11. Subpatterns that are repeated with specific counts have to be replicated in
1369 the compiled pattern. The size of memory for this was computed from the
1370 length of the subpattern and the repeat count. The latter is limited to
1371 65535, but there was no limit on the former, meaning that integer overflow
1372 could in principle occur. The compiled length of a repeated subpattern is
1373 now limited to 30,000 bytes in order to prevent this.
1375 12. Added the optional facility to have named substrings with the same name.
1377 13. Added the ability to use a named substring as a condition, using the
1378 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1379 are numbers (not recommended). Forward references are permitted.
1381 14. Added forward references in named backreferences (if you see what I mean).
1383 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1384 pattern could run off the end of the subject. For example, the pattern
1385 "(?s)(.{1,5})"8 did this with the subject "ab".
1387 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1388 PCRE_CASELESS was set when matching characters that were quantified with ?
1389 or *.
1391 17. A character class other than a single negated character that had a minimum
1392 but no maximum quantifier - for example [ab]{6,} - was not handled
1393 correctly by pce_dfa_exec(). It would match only one character.
1395 18. A valid (though odd) pattern that looked like a POSIX character
1396 class but used an invalid character after [ (for example [[,abc,]]) caused
1397 pcre_compile() to give the error "Failed: internal error: code overflow" or
1398 in some cases to crash with a glibc free() error. This could even happen if
1399 the pattern terminated after [[ but there just happened to be a sequence of
1400 letters, a binary zero, and a closing ] in the memory that followed.
1402 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1403 over the years. Originally (before any Unicode support), just the bottom 8
1404 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1405 output from "man perlunicode" includes this:
1407 The regular expression compiler produces polymorphic opcodes. That
1408 is, the pattern adapts to the data and automatically switches to
1409 the Unicode character scheme when presented with Unicode data--or
1410 instead uses a traditional byte scheme when presented with byte
1411 data.
1413 Sadly, a wide octal escape does not cause a switch, and in a string with
1414 no other multibyte characters, these octal escapes are treated as before.
1415 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1416 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1417 Unicode string.
1419 I have not perpetrated such confusion in PCRE. Up till now, it took just
1420 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1421 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1422 translate to the appropriate multibyte character.
1424 29. Applied some refactoring to reduce the number of warnings from Microsoft
1425 and Borland compilers. This has included removing the fudge introduced
1426 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1427 a warning about an unused variable.
1429 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1430 characters since release 4.0, because Perl (from release 5.004) does not.
1431 [Or at least, is documented not to: some releases seem to be in conflict
1432 with the documentation.] However, when a pattern was studied with
1433 pcre_study() and all its branches started with \s, PCRE still included VT
1434 as a possible starting character. Of course, this did no harm; it just
1435 caused an unnecessary match attempt.
1437 22. Removed a now-redundant internal flag bit that recorded the fact that case
1438 dependency changed within the pattern. This was once needed for "required
1439 byte" processing, but is no longer used. This recovers a now-scarce options
1440 bit. Also moved the least significant internal flag bit to the most-
1441 significant bit of the word, which was not previously used (hangover from
1442 the days when it was an int rather than a uint) to free up another bit for
1443 the future.
1445 23. Added support for CRLF line endings as well as CR and LF. As well as the
1446 default being selectable at build time, it can now be changed at runtime
1447 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1448 specify that it is scanning data with non-default line endings.
1450 24. Changed the definition of CXXLINK to make it agree with the definition of
1451 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1453 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1454 recursions. This makes a big different to stack usage for some patterns.
1456 26. If a subpattern containing a named recursion or subroutine reference such
1457 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1458 the space required for the compiled pattern went wrong and gave too small a
1459 value. Depending on the environment, this could lead to "Failed: internal
1460 error: code overflow at offset 49" or "glibc detected double free or
1461 corruption" errors.
1463 27. Applied patches from Google (a) to support the new newline modes and (b) to
1464 advance over multibyte UTF-8 characters in GlobalReplace.
1466 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1467 difference for some implementation of PCRE in some Windows version.
1469 29. Added some extra testing facilities to pcretest:
1471 \q<number> in a data line sets the "match limit" value
1472 \Q<number> in a data line sets the "match recursion limt" value
1473 -S <number> sets the stack size, where <number> is in megabytes
1475 The -S option isn't available for Windows.
1478 Version 6.6 06-Feb-06
1479 ---------------------
1481 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1482 in pcreposix.h. I have copied the definition from pcre.h.
1484 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1485 because pcre.h is no longer a built file.
1487 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1488 not normally included in the compiled code.
1491 Version 6.5 01-Feb-06
1492 ---------------------
1494 1. When using the partial match feature with pcre_dfa_exec(), it was not
1495 anchoring the second and subsequent partial matches at the new starting
1496 point. This could lead to incorrect results. For example, with the pattern
1497 /1234/, partially matching against "123" and then "a4" gave a match.
1499 2. Changes to pcregrep:
1501 (a) All non-match returns from pcre_exec() were being treated as failures
1502 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1503 error message is output. Some extra information is given for the
1505 probably the only errors that are likely to be caused by users (by
1506 specifying a regex that has nested indefinite repeats, for instance).
1507 If there are more than 20 of these errors, pcregrep is abandoned.
1509 (b) A binary zero was treated as data while matching, but terminated the
1510 output line if it was written out. This has been fixed: binary zeroes
1511 are now no different to any other data bytes.
1513 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1514 used to set a locale for matching. The --locale=xxxx long option has
1515 been added (no short equivalent) to specify a locale explicitly on the
1516 pcregrep command, overriding the environment variables.
1518 (d) When -B was used with -n, some line numbers in the output were one less
1519 than they should have been.
1521 (e) Added the -o (--only-matching) option.
1523 (f) If -A or -C was used with -c (count only), some lines of context were
1524 accidentally printed for the final match.
1526 (g) Added the -H (--with-filename) option.
1528 (h) The combination of options -rh failed to suppress file names for files
1529 that were found from directory arguments.
1531 (i) Added the -D (--devices) and -d (--directories) options.
1533 (j) Added the -F (--fixed-strings) option.
1535 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1537 (l) Added the --colo(u)r option.
1539 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1540 is not present by default.
1542 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1543 items such as (?R) or (?1), when the recursion could match a number of
1544 alternatives. If it matched one of the alternatives, but subsequently,
1545 outside the recursion, there was a failure, the code tried to back up into
1546 the recursion. However, because of the way PCRE is implemented, this is not
1547 possible, and the result was an incorrect result from the match.
1549 In order to prevent this happening, the specification of recursion has
1550 been changed so that all such subpatterns are automatically treated as
1551 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1553 4. I had overlooked the fact that, in some locales, there are characters for
1554 which isalpha() is true but neither isupper() nor islower() are true. In
1555 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1556 and ordfeminine) are like this. This affected the treatment of \w and \W
1557 when they appeared in character classes, but not when they appeared outside
1558 a character class. The bit map for "word" characters is now created
1559 separately from the results of isalnum() instead of just taking it from the
1560 upper, lower, and digit maps. (Plus the underscore character, of course.)
1562 5. The above bug also affected the handling of POSIX character classes such as
1563 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1564 permanent tables. Instead, the bit maps for such a class were previously
1565 created as the appropriate unions of the upper, lower, and digit bitmaps.
1566 Now they are created by subtraction from the [[:word:]] class, which has
1567 its own bitmap.
1569 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1570 It is created by subtracting the vertical space characters (\x09, \x0a,
1571 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1572 subtraction was done in the overall bitmap for a character class, meaning
1573 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1574 be recognized. This bug has been fixed.
1576 7. Patches from the folks at Google:
1578 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1579 real life, but is still worth protecting against".
1581 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1582 regular expressions".
1584 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1585 have it.
1587 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1588 "configure" and the latter not, in order to fix a problem somebody had
1589 with compiling the Arg class on HP-UX.
1591 (e) Improve the error-handling of the C++ wrapper a little bit.
1593 (f) New tests for checking recursion limiting.
1595 8. The pcre_memmove() function, which is used only if the environment does not
1596 have a standard memmove() function (and is therefore rarely compiled),
1597 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1598 returning a result (though PCRE never actually uses the result).
1600 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1601 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1602 returned instead of calling malloc() with an overflowing number that would
1603 most likely cause subsequent chaos.
1605 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1607 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1608 with this option is matched, the nmatch and pmatch options of regexec() are
1609 ignored.
1611 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1612 provided in case anyone wants to the the POSIX interface with UTF-8
1613 strings.
1615 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1616 C++ linking (needed for some HP-UX environments).
1618 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1619 (unused parameter) and in the pcre_printint() function (omitted "default"
1620 switch label when the default is to do nothing).
1622 15. Added some code to make it possible, when PCRE is compiled as a C++
1623 library, to replace subject pointers for pcre_exec() with a smart pointer
1624 class, thus making it possible to process discontinuous strings.
1626 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1627 much the same function. They were added by different people who were trying
1628 to make PCRE easy to compile on non-Unix systems. It has been suggested
1629 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1630 for compiling on Windows systems. I have therefore replaced it with
1631 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1632 defaults to "extern" for C or "extern C" for C++, which works fine on
1633 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1634 SCOPE with something explicit in config.h. In addition:
1636 (a) pcreposix.h still had just "extern" instead of either of these macros;
1637 I have replaced it with PCRE_DATA_SCOPE.
1639 (b) Functions such as _pcre_xclass(), which are internal to the library,
1640 but external in the C sense, all had PCRE_EXPORT in their definitions.
1641 This is apparently wrong for the Windows case, so I have removed it.
1642 (It makes no difference on Unix-like systems.)
1644 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1645 of recursive calls to match(). This is different to MATCH_LIMIT because
1646 that limits the total number of calls to match(), not all of which increase
1647 the depth of recursion. Limiting the recursion depth limits the amount of
1648 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1649 when PCRE is compiled, and changed at run time. A patch from Google adds
1650 this functionality to the C++ interface.
1652 18. Changes to the handling of Unicode character properties:
1654 (a) Updated the table to Unicode 4.1.0.
1656 (b) Recognize characters that are not in the table as "Cn" (undefined).
1658 (c) I revised the way the table is implemented to a much improved format
1659 which includes recognition of ranges. It now supports the ranges that
1660 are defined in UnicodeData.txt, and it also amalgamates other
1661 characters into ranges. This has reduced the number of entries in the
1662 table from around 16,000 to around 3,000, thus reducing its size
1663 considerably. I realized I did not need to use a tree structure after
1664 all - a binary chop search is just as efficient. Having reduced the
1665 number of entries, I extended their size from 6 bytes to 8 bytes to
1666 allow for more data.
1668 (d) Added support for Unicode script names via properties such as \p{Han}.
1670 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1671 matching that character.
1673 20. When matching a repeated Unicode property with a minimum greater than zero,
1674 (for example \pL{2,}), PCRE could look past the end of the subject if it
1675 reached it while seeking the minimum number of characters. This could
1676 happen only if some of the characters were more than one byte long, because
1677 there is a check for at least the minimum number of bytes.
1679 21. Refactored the implementation of \p and \P so as to be more general, to
1680 allow for more different types of property in future. This has changed the
1681 compiled form incompatibly. Anybody with saved compiled patterns that use
1682 \p or \P will have to recompile them.
1684 22. Added "Any" and "L&" to the supported property types.
1686 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1687 but give a compile time error if the value is greater than 0xff.
1689 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1690 accidentally not being installed or uninstalled.
1692 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1693 made were to insert the current release number. This seemed silly, because
1694 it made things harder for people building PCRE on systems that don't run
1695 "configure". I have turned pcre.h into a distributed file, no longer built
1696 by "configure", with the version identification directly included. There is
1697 no longer a pcre.h.in file.
1699 However, this change necessitated a change to the pcre-config script as
1700 well. It is built from pcre-config.in, and one of the substitutions was the
1701 release number. I have updated configure.ac so that ./configure now finds
1702 the release number by grepping pcre.h.
1704 26. Added the ability to run the tests under valgrind.
1707 Version 6.4 05-Sep-05
1708 ---------------------
1710 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1711 "--" to be printed when multiple files were scanned, even when none of the
1712 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1713 consider it to be a bug, and have restored the previous behaviour.
1715 2. A couple of code tidies to get rid of compiler warnings.
1717 3. The pcretest program used to cheat by referring to symbols in the library
1718 whose names begin with _pcre_. These are internal symbols that are not
1719 really supposed to be visible externally, and in some environments it is
1720 possible to suppress them. The cheating is now confined to including
1721 certain files from the library's source, which is a bit cleaner.
1723 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1724 file's purpose clearer.
1726 5. Reorganized pcre_ucp_findchar().
1729 Version 6.3 15-Aug-05
1730 ---------------------
1732 1. The file libpcre.pc.in did not have general read permission in the tarball.
1734 2. There were some problems when building without C++ support:
1736 (a) If C++ support was not built, "make install" and "make test" still
1737 tried to test it.
1739 (b) There were problems when the value of CXX was explicitly set. Some
1740 changes have been made to try to fix these, and ...
1742 (c) --disable-cpp can now be used to explicitly disable C++ support.
1744 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1745 backslash in a target when C++ was disabled. This confuses some
1746 versions of "make", apparently. Using an intermediate variable solves
1747 this. (Same for CPP_LOBJ.)
1749 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1750 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1751 necessary on certain architectures.
1753 4. Added a setting of -export-symbols-regex to the link command to remove
1754 those symbols that are exported in the C sense, but actually are local
1755 within the library, and not documented. Their names all begin with
1756 "_pcre_". This is not a perfect job, because (a) we have to except some
1757 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1758 available (and never for static libraries). I have made a note to try to
1759 find a way round (a) in the future.
1762 Version 6.2 01-Aug-05
1763 ---------------------
1765 1. There was no test for integer overflow of quantifier values. A construction
1766 such as {1111111111111111} would give undefined results. What is worse, if
1767 a minimum quantifier for a parenthesized subpattern overflowed and became
1768 negative, the calculation of the memory size went wrong. This could have
1769 led to memory overwriting.
1771 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1773 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1774 operating environments where this matters.
1776 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1777 PCRE options from within the C++ wrapper.
1779 5. Named capturing subpatterns were not being correctly counted when a pattern
1780 was compiled. This caused two problems: (a) If there were more than 100
1781 such subpatterns, the calculation of the memory needed for the whole
1782 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1783 back references of the form \12, where the number was greater than 9, were
1784 not recognized as back references, even though there were sufficient
1785 previous subpatterns.
1787 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1788 versions of gcc, e.g. 2.95.4.
1791 Version 6.1 21-Jun-05
1792 ---------------------
1794 1. There was one reference to the variable "posix" in pcretest.c that was not
1795 surrounded by "#if !defined NOPOSIX".
1797 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1798 the cross-check on the old pcre_info() function, for the benefit of the
1799 cut-down version of PCRE that is currently imported into Exim.
1801 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1802 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1803 patterns that start (?i) but I don't think that matters. The use of (?i) is
1804 just an example; this all applies to the other options as well.
1806 4. Since libtool seems to echo the compile commands it is issuing, the output
1807 from "make" can be reduced a bit by putting "@" in front of each libtool
1808 compile command.
1810 5. Patch from the folks at Google for configure.in to be a bit more thorough
1811 in checking for a suitable C++ installation before trying to compile the
1812 C++ stuff. This should fix a reported problem when a compiler was present,
1813 but no suitable headers.
1815 6. The man pages all had just "PCRE" as their title. I have changed them to
1816 be the relevant file name. I have also arranged that these names are
1817 retained in the file doc/pcre.txt, which is a concatenation in text format
1818 of all the man pages except the little individual ones for each function.
1820 7. The NON-UNIX-USE file had not been updated for the different set of source
1821 files that come with release 6. I also added a few comments about the C++
1822 wrapper.
1825 Version 6.0 07-Jun-05
1826 ---------------------
1828 1. Some minor internal re-organization to help with my DFA experiments.
1830 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1831 didn't matter for the library itself when fully configured, but did matter
1832 when compiling without UCP support, or within Exim, where the ucp files are
1833 not imported.
1835 3. Refactoring of the library code to split up the various functions into
1836 different source modules. The addition of the new DFA matching code (see
1837 below) to a single monolithic source would have made it really too
1838 unwieldy, quite apart from causing all the code to be include in a
1839 statically linked application, when only some functions are used. This is
1840 relevant even without the DFA addition now that patterns can be compiled in
1841 one application and matched in another.
1843 The downside of splitting up is that there have to be some external
1844 functions and data tables that are used internally in different modules of
1845 the library but which are not part of the API. These have all had their
1846 names changed to start with "_pcre_" so that they are unlikely to clash
1847 with other external names.
1849 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1850 a different (DFA) algorithm. Although it is slower than the original
1851 function, it does have some advantages for certain types of matching
1852 problem.
1854 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1855 including restarting after a partial match.
1857 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1858 defined when compiling for Windows was sent to me. I have put it into the
1859 code, though I have no means of testing or verifying it.
1861 7. Added the pcre_refcount() auxiliary function.
1863 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1864 match before or at the first newline in the subject string. In pcretest,
1865 the /f option on a pattern can be used to set this.
1867 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1868 would behave wrongly. This has been present in PCRE since release 4.0.
1870 10. A number of changes to the pcregrep command:
1872 (a) Refactored how -x works; insert ^(...)$ instead of setting
1873 PCRE_ANCHORED and checking the length, in preparation for adding
1874 something similar for -w.
1876 (b) Added the -w (match as a word) option.
1878 (c) Refactored the way lines are read and buffered so as to have more
1879 than one at a time available.
1881 (d) Implemented a pcregrep test script.
1883 (e) Added the -M (multiline match) option. This allows patterns to match
1884 over several lines of the subject. The buffering ensures that at least
1885 8K, or the rest of the document (whichever is the shorter) is available
1886 for matching (and similarly the previous 8K for lookbehind assertions).
1888 (f) Changed the --help output so that it now says
1890 -w, --word-regex(p)
1892 instead of two lines, one with "regex" and the other with "regexp"
1893 because that confused at least one person since the short forms are the
1894 same. (This required a bit of code, as the output is generated
1895 automatically from a table. It wasn't just a text change.)
1897 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1898 option but starts with a hyphen. Could be a pattern or a path name
1899 starting with a hyphen, for instance.
1901 (h) "-" can be given as a file name to represent stdin.
1903 (i) When file names are being printed, "(standard input)" is used for
1904 the standard input, for compatibility with GNU grep. Previously
1905 "<stdin>" was used.
1907 (j) The option --label=xxx can be used to supply a name to be used for
1908 stdin when file names are being printed. There is no short form.
1910 (k) Re-factored the options decoding logic because we are going to add
1911 two more options that take data. Such options can now be given in four
1912 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1914 (l) Added the -A, -B, and -C options for requesting that lines of context
1915 around matches be printed.
1917 (m) Added the -L option to print the names of files that do not contain
1918 any matching lines, that is, the complement of -l.
1920 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1921 continue to scan other files.
1923 (o) The -s option was incorrectly implemented. For compatibility with other
1924 greps, it now suppresses the error message for a non-existent or non-
1925 accessible file (but not the return code). There is a new option called
1926 -q that suppresses the output of matching lines, which was what -s was
1927 previously doing.
1929 (p) Added --include and --exclude options to specify files for inclusion
1930 and exclusion when recursing.
1932 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1933 Hopefully, it now does.
1935 12. Missing cast in pcre_study().
1937 13. Added an "uninstall" target to the makefile.
1939 14. Replaced "extern" in the function prototypes in Makefile.in with
1940 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1941 world, but is set differently for Windows.
1943 15. Added a second compiling function called pcre_compile2(). The only
1944 difference is that it has an extra argument, which is a pointer to an
1945 integer error code. When there is a compile-time failure, this is set
1946 non-zero, in addition to the error test pointer being set to point to an
1947 error message. The new argument may be NULL if no error number is required
1948 (but then you may as well call pcre_compile(), which is now just a
1949 wrapper). This facility is provided because some applications need a
1950 numeric error indication, but it has also enabled me to tidy up the way
1951 compile-time errors are handled in the POSIX wrapper.
1953 16. Added VPATH=.libs to the makefile; this should help when building with one
1954 prefix path and installing with another. (Or so I'm told by someone who
1955 knows more about this stuff than I do.)
1957 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1958 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1959 match everything, including newlines. This is not POSIX-compatible, but
1960 somebody wanted the feature. From pcretest it can be activated by using
1961 both the P and the s flags.
1963 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1965 19. libpcre.pc was being incorrectly installed as executable.
1967 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1968 it now also looks for '\r' so that it will work unmodified on Windows.
1970 21. Added Google's contributed C++ wrapper to the distribution.
1972 22. Added some untidy missing memory free() calls in pcretest, to keep
1973 Electric Fence happy when testing.
1977 Version 5.0 13-Sep-04
1978 ---------------------
1980 1. Internal change: literal characters are no longer packed up into items
1981 containing multiple characters in a single byte-string. Each character
1982 is now matched using a separate opcode. However, there may be more than one
1983 byte in the character in UTF-8 mode.
1985 2. The pcre_callout_block structure has two new fields: pattern_position and
1986 next_item_length. These contain the offset in the pattern to the next match
1987 item, and its length, respectively.
1989 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1990 insertion of callouts before each pattern item. Added the /C option to
1991 pcretest to make use of this.
1993 4. On the advice of a Windows user, the lines
1995 #if defined(_WIN32) || defined(WIN32)
1996 _setmode( _fileno( stdout ), 0x8000 );
1997 #endif /* defined(_WIN32) || defined(WIN32) */
1999 have been added to the source of pcretest. This apparently does useful
2000 magic in relation to line terminators.
2002 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2003 for the benefit of those environments where the "b" makes a difference.
2005 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2006 to know about it. I have put a hack into configure.in that adds in code
2007 to set GCC=yes if CC=icc. This seems to end up at a point in the
2008 generated configure script that is early enough to affect the setting of
2009 compiler options, which is what is needed, but I have no means of testing
2010 whether it really works. (The user who reported this had patched the
2011 generated configure script, which of course I cannot do.)
2013 LATER: After change 22 below (new libtool files), the configure script
2014 seems to know about icc (and also ecc). Therefore, I have commented out
2015 this hack in configure.in.
2017 7. Added support for pkg-config (2 patches were sent in).
2019 8. Negated POSIX character classes that used a combination of internal tables
2020 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2021 [[:^ascii]]. Typically, they would match almost any characters. The other
2022 POSIX classes were not broken in this way.
2024 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2025 to find the match, as PCRE was deluded into thinking that the match had to
2026 start at the start point or following a newline. The same bug applied to
2027 patterns with negative forward assertions or any backward assertions
2028 preceding ".*" at the start, unless the pattern required a fixed first
2029 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2031 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2032 starting at the last subject character, bytes beyond the end of the subject
2033 string were read.
2035 11. Renamed the variable "class" as "classbits" to make life easier for C++
2036 users. (Previously there was a macro definition, but it apparently wasn't
2037 enough.)
2039 12. Added the new field "tables" to the extra data so that tables can be passed
2040 in at exec time, or the internal tables can be re-selected. This allows
2041 a compiled regex to be saved and re-used at a later time by a different
2042 program that might have everything at different addresses.
2044 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2045 -R library as well as a -L library.
2047 14. The debugging options of pcretest (-d on the command line or D on a
2048 pattern) showed incorrect output for anything following an extended class
2049 that contained multibyte characters and which was followed by a quantifier.
2051 15. Added optional support for general category Unicode character properties
2052 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2053 support. It adds about 90K to the size of the library. The meanings of the
2054 inbuilt class escapes such as \d and \s have NOT been changed.
2056 16. Updated pcredemo.c to include calls to free() to release the memory for the
2057 compiled pattern.
2059 17. The generated file chartables.c was being created in the source directory
2060 instead of in the building directory. This caused the build to fail if the
2061 source directory was different from the building directory, and was
2062 read-only.
2064 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2065 file. No doubt somebody will tell me if they don't make sense... Also added
2066 Dan Mooney's comments about building on OpenVMS.
2068 19. Added support for partial matching via the PCRE_PARTIAL option for
2069 pcre_exec() and the \P data escape in pcretest.
2071 20. Extended pcretest with 3 new pattern features:
2073 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2074 write the compiled pattern to the file whose name is "rest-of-line".
2075 This is a straight binary dump of the data, with the saved pointer to
2076 the character tables forced to be NULL. The study data, if any, is
2077 written too. After writing, pcretest reads a new pattern.
2079 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2080 compiled pattern from the given file. There must not be any
2081 occurrences of "<" in the file name (pretty unlikely); if there are,
2082 pcretest will instead treat the initial "<" as a pattern delimiter.
2083 After reading in the pattern, pcretest goes on to read data lines as
2084 usual.
2086 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2087 and 16-bit fields in a compiled pattern, to simulate a pattern that
2088 was compiled on a host of opposite endianness.
2090 21. The pcre-exec() function can now cope with patterns that were compiled on
2091 hosts of opposite endianness, with this restriction:
2093 As for any compiled expression that is saved and used later, the tables
2094 pointer field cannot be preserved; the extra_data field in the arguments
2095 to pcre_exec() should be used to pass in a tables address if a value
2096 other than the default internal tables were used at compile time.
2098 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2099 now diagnosed as an error. Previously, most of the time, a negative number
2100 would have been treated as zero, but if in addition "ovector" was passed as
2101 NULL, a crash could occur.
2103 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2104 new versions from the libtool 1.5 distribution (the last one is a copy of
2105 a file called libtool.m4). This seems to have fixed the need to patch
2106 "configure" to support Darwin 1.3 (which I used to do). However, I still
2107 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2108 workstation).
2110 24. Changed the PCRE licence to be the more standard "BSD" licence.
2113 Version 4.5 01-Dec-03
2114 ---------------------
2116 1. There has been some re-arrangement of the code for the match() function so
2117 that it can be compiled in a version that does not call itself recursively.
2118 Instead, it keeps those local variables that need separate instances for
2119 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2120 needs to "recurse". Keeping track of where control must go is done by means
2121 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2122 hide most of the details from the main code, and operates only if
2123 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2124 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2125 operating.
2127 To make it easier for callers to provide specially tailored get/free
2128 functions for this usage, two new functions, pcre_stack_malloc, and
2129 pcre_stack_free, are used. They are always called in strict stacking order,
2130 and the size of block requested is always the same.
2132 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2133 PCRE has been compiled to use the stack or the heap for recursion. The
2134 -C option of pcretest uses this to show which version is compiled.
2136 A new data escape \S, is added to pcretest; it causes the amounts of store
2137 obtained and freed by both kinds of malloc/free at match time to be added
2138 to the output.
2140 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2141 what's available on my current Linux desktop machine.
2143 3. When matching a UTF-8 string, the test for a valid string at the start has
2144 been extended. If start_offset is not zero, PCRE now checks that it points
2145 to a byte that is the start of a UTF-8 character. If not, it returns
2146 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2147 this is necessary because there may be backward assertions in the pattern.
2148 When matching the same subject several times, it may save resources to use
2149 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2151 4. The code for checking the validity of UTF-8 strings has been tightened so
2152 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2153 containing "overlong sequences".
2155 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2156 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2157 should have been "&", but it just so happened that all the cases this let
2158 through by mistake were picked up later in the function.
2160 6. I had used a variable called "isblank" - this is a C99 function, causing
2161 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2163 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2164 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2165 so the version is shown; (c) stop "make test" echoing "./RunTest".
2167 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2169 9. The prototype for memmove() for systems that don't have it was using
2170 size_t, but the inclusion of the header that defines size_t was later. I've
2171 moved the #includes for the C headers earlier to avoid this.
2173 10. Added some adjustments to the code to make it easier to compiler on certain
2174 special systems:
2176 (a) Some "const" qualifiers were missing.
2177 (b) Added the macro EXPORT before all exported functions; by default this
2178 is defined to be empty.
2179 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2180 that it reads its output file name as an argument instead of writing
2181 to the standard output and assuming this can be redirected.
2183 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2184 class containing characters with values greater than 255, PCRE compilation
2185 went into a loop.
2187 12. A recursive reference to a subpattern that was within another subpattern
2188 that had a minimum quantifier of zero caused PCRE to crash. For example,
2189 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2190 recursion. If the recursively-called subpattern itself had a zero repeat,
2191 that was OK.
2193 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2194 buffer into which it was copied (for escape processing) was still set at
2195 1024, so long lines caused crashes.
2197 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2198 "internal error: code overflow...". This applied to any character class
2199 that was followed by a possessive quantifier.
2201 15. Modified the Makefile to add libpcre.la as a prerequisite for
2202 libpcreposix.la because I was told this is needed for a parallel build to
2203 work.
2205 16. If a pattern that contained .* following optional items at the start was
2206 studied, the wrong optimizing data was generated, leading to matching
2207 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2208 matching string must start with a or b or c. The correct conclusion for
2209 this pattern is that a match can start with any character.
2212 Version 4.4 13-Aug-03
2213 ---------------------
2215 1. In UTF-8 mode, a character class containing characters with values between
2216 127 and 255 was not handled correctly if the compiled pattern was studied.
2217 In fixing this, I have also improved the studying algorithm for such
2218 classes (slightly).
2220 2. Three internal functions had redundant arguments passed to them. Removal
2221 might give a very teeny performance improvement.
2223 3. Documentation bug: the value of the capture_top field in a callout is *one
2224 more than* the number of the hightest numbered captured substring.
2226 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2227 in incorrectly linking with a previously installed version. They now link
2228 explicitly with libpcre.la.
2230 5. configure.in no longer needs to recognize Cygwin specially.
2232 6. A problem in pcre.in for Windows platforms is fixed.
2234 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2235 pcretest, it used to include the size of the study block as part of its
2236 output. Unfortunately, the structure contains a field that has a different
2237 size on different hardware architectures. This meant that the tests that
2238 showed this size failed. As the block is currently always of a fixed size,
2239 this information isn't actually particularly useful in pcretest output, so
2240 I have just removed it.
2242 8. Three pre-processor statements accidentally did not start in column 1.
2243 Sadly, there are *still* compilers around that complain, even though
2244 standard C has not required this for well over a decade. Sigh.
2246 9. In pcretest, the code for checking callouts passed small integers in the
2247 callout_data field, which is a void * field. However, some picky compilers
2248 complained about the casts involved for this on 64-bit systems. Now
2249 pcretest passes the address of the small integer instead, which should get
2250 rid of the warnings.
2252 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2253 both compile and run time, and gives an error if an invalid UTF-8 sequence
2254 is found. There is a option for disabling this check in cases where the
2255 string is known to be correct and/or the maximum performance is wanted.
2257 11. In response to a bug report, I changed one line in Makefile.in from
2259 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2260 to
2261 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2263 to look similar to other lines, but I have no way of telling whether this
2264 is the right thing to do, as I do not use Windows. No doubt I'll get told
2265 if it's wrong...
2268 Version 4.3 21-May-03
2269 ---------------------
2271 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2272 Makefile.
2274 2. Some refactoring to improve the quality of the code:
2276 (i) The utf8_table... variables are now declared "const".
2278 (ii) The code for \cx, which used the "case flipping" table to upper case
2279 lower case letters, now just substracts 32. This is ASCII-specific,
2280 but the whole concept of \cx is ASCII-specific, so it seems
2281 reasonable.
2283 (iii) PCRE was using its character types table to recognize decimal and
2284 hexadecimal digits in the pattern. This is silly, because it handles
2285 only 0-9, a-f, and A-F, but the character types table is locale-
2286 specific, which means strange things might happen. A private
2287 table is now used for this - though it costs 256 bytes, a table is
2288 much faster than multiple explicit tests. Of course, the standard
2289 character types table is still used for matching digits in subject
2290 strings against \d.
2292 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2293 ending in _t are). So I've renamed it as ESC_tee.
2295 3. The first argument for regexec() in the POSIX wrapper should have been
2296 defined as "const".
2298 4. Changed pcretest to use malloc() for its buffers so that they can be
2299 Electric Fenced for debugging.
2301 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2302 to read one or more bytes before the start of the subject string. Often this
2303 had no effect on PCRE's behaviour, but in some circumstances it could
2304 provoke a segmentation fault.
2306 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2307 to try to read one or more bytes before the start of the subject string.
2309 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2310 UTF-8 support could misbehave in various ways if the subject string
2311 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2312 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2313 back over UTF-8 characters.)
2316 Version 4.2 14-Apr-03
2317 ---------------------
2319 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2321 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2322 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2323 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2324 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2325 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2327 Note: automatic setting of the BUILD variables is not yet working
2328 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2329 compile-time but not at link-time
2330 [LINK]: use for linking executables only
2331 make different versions for Windows and non-Windows
2332 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2333 libraries
2334 [LINK_FOR_BUILD]: new variable
2335 [OBJEXT]: use throughout
2336 [EXEEXT]: use throughout
2337 <winshared>: new target
2338 <wininstall>: new target
2339 <dftables.o>: use native compiler
2340 <dftables>: use native linker
2341 <install>: handle Windows platform correctly
2342 <clean>: ditto
2343 <check>: ditto
2344 copy DLL to top builddir before testing
2346 As part of these changes, -no-undefined was removed again. This was reported
2347 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2348 in any case.
2350 3. Some tidies to get rid of compiler warnings:
2352 . In the match_data structure, match_limit was an unsigned long int, whereas
2353 match_call_count was an int. I've made them both unsigned long ints.
2355 . In pcretest the fact that a const uschar * doesn't automatically cast to
2356 a void * provoked a warning.
2358 . Turning on some more compiler warnings threw up some "shadow" variables
2359 and a few more missing casts.
2361 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2362 option, a class that contained a single character with a value between 128
2363 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2365 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2366 option, a class that contained several characters, but with at least one
2367 whose value was between 128 and 255 caused PCRE to crash.
2370 Version 4.1 12-Mar-03
2371 ---------------------
2373 1. Compiling with gcc -pedantic found a couple of places where casts were
2374 needed, and a string in dftables.c that was longer than standard compilers are
2375 required to support.
2377 2. Compiling with Sun's compiler found a few more places where the code could
2378 be tidied up in order to avoid warnings.
2380 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2381 first of these names is deprecated in the latest Autoconf in favour of the name
2382 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2383 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2384 analogy I have changed it to CFLAGS_FOR_BUILD.
2386 4. Added -no-undefined to the linking command in the Makefile, because this is
2387 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2388 linking step for the pcreposix library.
2390 5. PCRE was failing to diagnose the case of two named groups with the same
2391 name.
2393 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2394 literal character that is needed in the subject for a match, and scans along to
2395 ensure that it is present before embarking on the full matching process. This
2396 saves time in cases of nested unlimited repeats that are never going to match.
2397 Problem: the scan can take a lot of time if the subject is very long (e.g.
2398 megabytes), thus penalizing straightforward matches. It is now done only if the
2399 amount of subject to be scanned is less than 1000 bytes.
2401 7. A lesser problem with the same optimization is that it was recording the
2402 first character of an anchored pattern as "needed", thus provoking a search
2403 right along the subject, even when the first match of the pattern was going to
2404 fail. The "needed" character is now not set for anchored patterns, unless it
2405 follows something in the pattern that is of non-fixed length. Thus, it still
2406 fulfils its original purpose of finding quick non-matches in cases of nested
2407 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2410 Version 4.0 17-Feb-03
2411 ---------------------
2413 1. If a comment in an extended regex that started immediately after a meta-item
2414 extended to the end of string, PCRE compiled incorrect data. This could lead to
2415 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2417 2. Moved to autoconf 2.53 and libtool 1.4.2.
2419 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2420 the special perltest8 script is no longer needed - all the tests can be run
2421 from a single perltest script.
2423 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2424 by \s. It has now been removed in PCRE. This means it isn't recognized as
2425 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2426 class [:space:] *does* include VT, thereby creating a mess.
2428 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2429 space and tab.
2431 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2432 its new features into the main test script, reducing the number of scripts.
2434 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2435 were backward compatible, and made the (?i) apply to the whole pattern, as if
2436 /i were given. Now it behaves more logically, and applies the option setting
2437 only to what follows. PCRE has been changed to follow suit. However, if it
2438 finds options settings right at the start of the pattern, it extracts them into
2439 the global options, as before. Thus, they show up in the info data.
2441 8. Added support for the \Q...\E escape sequence. Characters in between are
2442 treated as literals. This is slightly different from Perl in that $ and @ are
2443 also handled as literals inside the quotes. In Perl, they will cause variable
2444 interpolation. Note the following examples:
2446 Pattern PCRE matches Perl matches
2448 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2449 \Qabc\$xyz\E abc\$xyz abc\$xyz
2450 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2452 For compatibility with Perl, \Q...\E sequences are recognized inside character
2453 classes as well as outside them.
2455 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2456 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2457 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2458 signed/unsigned warnings.
2460 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2461 option for pcretest, so I've replaced it by a simple function that does just
2462 that job.
2464 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2465 "pcregrep -".
2467 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2468 Java package. This provides some syntactic sugar for simple cases of what my
2469 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2470 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2471 item, you can use this simplified notation. Note that only makes sense with
2472 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2473 greediness, whatever the setting of the PCRE_UNGREEDY option.
2475 13. A change of greediness default within a pattern was not taking effect at
2476 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2477 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2478 was abstracted outside.
2480 14. PCRE now supports the \G assertion. It is true when the current matching
2481 position is at the start point of the match. This differs from \A when the
2482 starting offset is non-zero. Used with the /g option of pcretest (or similar
2483 code), it works in the same way as it does for Perl's /g option. If all
2484 alternatives of a regex begin with \G, the expression is anchored to the start
2485 match position, and the "anchored" flag is set in the compiled expression.
2487 15. Some bugs concerning the handling of certain option changes within patterns
2488 have been fixed. These applied to options other than (?ims). For example,
2489 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2490 been the other way round. Some of this was related to change 7 above.
2492 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2493 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2494 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2495 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2497 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2498 mode. Unlike ".", it always matches newline, whatever the setting of
2499 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2500 assertions. Perl allows it, but it doesn't (in general) work because it can't
2501 calculate the length of the lookbehind. At least, that's the case for Perl
2502 5.8.0 - I've been told they are going to document that it doesn't work in
2503 future.
2505 18. Added an error diagnosis for escapes that PCRE does not support: these are
2506 \L, \l, \N, \P, \p, \U, \u, and \X.
2508 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2509 reading past the end of the pattern in cases such as /[abcd/.
2511 20. PCRE was getting more memory than necessary for patterns with classes that
2512 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2514 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2515 compiling PCRE for use with Virtual Pascal.
2517 22. Small fix to the Makefile to make it work properly if the build is done
2518 outside the source tree.
2520 23. Added a new extension: a condition to go with recursion. If a conditional
2521 subpattern starts with (?(R) the "true" branch is used if recursion has
2522 happened, whereas the "false" branch is used only at the top level.
2524 24. When there was a very long string of literal characters (over 255 bytes
2525 without UTF support, over 250 bytes with UTF support), the computation of how
2526 much memory was required could be incorrect, leading to segfaults or other
2527 strange effects.
2529 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2530 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2531 there was a subsequent back reference to those brackets. This meant that, for
2532 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2533 possible to check for precisely this case. All we can do is abandon the
2534 optimization if .* occurs inside capturing brackets when there are any back
2535 references whatsoever. (See below for a better fix that came later.)
2537 26. The handling of the optimization for finding the first character of a
2538 non-anchored pattern, and for finding a character that is required later in the
2539 match were failing in some cases. This didn't break the matching; it just
2540 failed to optimize when it could. The way this is done has been re-implemented.
2542 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2544 28. Added a new feature that provides some of the functionality that Perl
2545 provides with (?{...}). The facility is termed a "callout". The way it is done
2546 in PCRE is for the caller to provide an optional function, by setting
2547 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2548 global variable. By default it is unset, which disables all calling out. To get
2549 the function called, the regex must include (?C) at appropriate points. This
2550 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2551 This provides a means of identifying different callout points. When PCRE
2552 reaches such a point in the regex, if pcre_callout has been set, the external
2553 function is called. It is provided with data in a structure called
2554 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2555 matching continues; if it returns a non-zero value, the match at the current
2556 point fails. However, backtracking will occur if possible. [This was changed
2557 later and other features added - see item 49 below.]
2559 29. pcretest is upgraded to test the callout functionality. It provides a
2560 callout function that displays information. By default, it shows the start of
2561 the match and the current position in the text. There are some new data escapes
2562 to vary what happens:
2564 \C+ in addition, show current contents of captured substrings
2565 \C- do not supply a callout function
2566 \C!n return 1 when callout number n is reached
2567 \C!n!m return 1 when callout number n is reached for the mth time
2569 30. If pcregrep was called with the -l option and just a single file name, it
2570 output "<stdin>" if a match was found, instead of the file name.
2572 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2573 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2574 pcre_exec(). This saves a malloc/free per call. The default value of
2575 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2576 when configuring.
2578 32. The default maximum size of a compiled pattern is 64K. There have been a
2579 few cases of people hitting this limit. The code now uses macros to handle the
2580 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2581 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2582 configuring. Tests 2 and 5 work only with 2-byte links because they output
2583 debugging information about compiled patterns.
2585 33. Internal code re-arrangements:
2587 (a) Moved the debugging function for printing out a compiled regex into
2588 its own source file (printint.c) and used #include to pull it into
2589 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2590 separate copies.
2592 (b) Defined the list of op-code names for debugging as a macro in
2593 internal.h so that it is next to the definition of the opcodes.
2595 (c) Defined a table of op-code lengths for simpler skipping along compiled
2596 code. This is again a macro in internal.h so that it is next to the
2597 definition of the opcodes.
2599 34. Added support for recursive calls to individual subpatterns, along the
2600 lines of Robin Houston's patch (but implemented somewhat differently).
2602 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2603 allow it to read and process whole directories in Win32. This code was
2604 contributed by Lionel Fourquaux; it has not been tested by me.
2606 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2607 used to name a group. Names consist of alphanumerics and underscores, and must
2608 be unique. Back references use the syntax (?P=name) and recursive calls use
2609 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2610 numbers. The function pcre_fullinfo() can be used after compilation to extract
2611 a name/number map. There are three relevant calls:
2613 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2614 PCRE_INFO_NAMECOUNT yields the number of entries
2615 PCRE_INFO_NAMETABLE yields a pointer to the map.
2617 The map is a vector of fixed-size entries. The size of each entry depends on
2618 the length of the longest name used. The first two bytes of each entry are the
2619 group number, most significant byte first. There follows the corresponding
2620 name, zero terminated. The names are in alphabetical order.
2622 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2623 case instead of 255. Making it the same both with and without UTF-8 support
2624 means that the same test output works with both.
2626 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2627 calling malloc() with a zero argument.
2629 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2630 optimization. I've improved things by keeping a bitmap of backreferences with
2631 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2632 fact referenced, the optimization can be applied. It is unlikely that a
2633 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2634 the match to follow \n) will appear inside brackets with a number greater than
2635 31, but if it does, any back reference > 31 suppresses the optimization.
2637 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2638 of disabling numbered capturing parentheses. Any opening parenthesis that is
2639 not followed by ? behaves as if it were followed by ?: but named parentheses
2640 can still be used for capturing (and they will acquire numbers in the usual
2641 way).
2643 41. Redesigned the return codes from the match() function into yes/no/error so
2644 that errors can be passed back from deep inside the nested calls. A malloc
2645 failure while inside a recursive subpattern call now causes the
2646 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2648 42. It is now possible to set a limit on the number of times the match()
2649 function is called in a call to pcre_exec(). This facility makes it possible to
2650 limit the amount of recursion and backtracking, though not in a directly
2651 obvious way, because the match() function is used in a number of different
2652 circumstances. The count starts from zero for each position in the subject
2653 string (for non-anchored patterns). The default limit is, for compatibility, a
2654 large number, namely 10 000 000. You can change this in two ways:
2656 (a) When configuring PCRE before making, you can use --with-match-limit=n
2657 to set a default value for the compiled library.
2659 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2660 a different value is set. See 45 below.
2662 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2664 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2665 of things that can be changed at compile time. The first argument specifies
2666 what is wanted and the second points to where the information is to be placed.
2667 The current list of available information is:
2671 The output is an integer that is set to one if UTF-8 support is available;
2672 otherwise it is set to zero.
2676 The output is an integer that it set to the value of the code that is used for
2677 newline. It is either LF (10) or CR (13).
2681 The output is an integer that contains the number of bytes used for internal
2682 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2686 The output is an integer that contains the threshold above which the POSIX
2687 interface uses malloc() for output vectors. See item 31 above.
2691 The output is an unsigned integer that contains the default limit of the number
2692 of match() calls in a pcre_exec() execution. See 42 above.
2694 44. pcretest has been upgraded by the addition of the -C option. This causes it
2695 to extract all the available output from the new pcre_config() function, and to
2696 output it. The program then exits immediately.
2698 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2699 order to support additional features. One way would have been to define
2700 pcre_exec2() (for example) with extra arguments, but this would not have been
2701 extensible, and would also have required all calls to the original function to
2702 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2703 is used for passing in "extra" data from pcre_study().
2705 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2706 contains the following fields:
2708 flags a bitmap indicating which of the following fields are set
2709 study_data opaque data from pcre_study()
2710 match_limit a way of specifying a limit on match() calls for a specific
2711 call to pcre_exec()
2712 callout_data data for callouts (see 49 below)
2714 The flag bits are also defined in pcre.h, and are
2720 The pcre_study() function now returns one of these new pcre_extra blocks, with
2721 the actual study data pointed to by the study_data field, and the
2722 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2723 before. That is, this change is entirely upwards-compatible and requires no
2724 change to existing code.
2726 If you want to pass in additional data to pcre_exec(), you can either place it
2727 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2728 block.
2730 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2731 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2732 times with different match limits, until it finds the minimum value needed for
2733 pcre_exec() to complete. The value is then output. This can be instructive; for
2734 most simple matches the number is quite small, but for pathological cases it
2735 gets very large very quickly.
2737 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2738 returns the size of the data block pointed to by the study_data field in a
2739 pcre_extra block, that is, the value that was passed as the argument to
2740 pcre_malloc() when PCRE was getting memory in which to place the information
2741 created by pcre_study(). The fourth argument should point to a size_t variable.
2742 pcretest has been extended so that this information is shown after a successful
2743 pcre_study() call when information about the compiled regex is being displayed.
2745 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2746 because what follows is always an absolute path. (Later: it turns out that this
2747 is more than cosmetic for MinGW, because it doesn't like empty path
2748 components.)
2750 49. Some changes have been made to the callout feature (see 28 above):
2752 (i) A callout function now has three choices for what it returns:
2754 0 => success, carry on matching
2755 > 0 => failure at this point, but backtrack if possible
2756 < 0 => serious error, return this value from pcre_exec()
2758 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2759 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2760 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2761 use by callout functions. It will never be used by PCRE itself.
2763 (ii) The pcre_extra structure (see 45 above) has a void * field called
2764 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2765 pcre_callout_block structure has a field of the same name. The contents of
2766 the field passed in the pcre_extra structure are passed to the callout
2767 function in the corresponding field in the callout block. This makes it
2768 easier to use the same callout-containing regex from multiple threads. For
2769 testing, the pcretest program has a new data escape
2771 \C*n pass the number n (may be negative) as callout_data
2773 If the callout function in pcretest receives a non-zero value as
2774 callout_data, it returns that value.
2776 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2777 there were some redundant $(CFLAGS) in commands that are now specified as
2778 $(LINK), which already includes $(CFLAGS).
2780 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2781 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2782 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2783 one-byte characters throughout. Note that case-insensitive matching applies
2784 only to characters whose values are less than 256. PCRE doesn't support the
2785 notion of cases for higher-valued characters.
2787 (i) A character class whose characters are all within 0-255 is handled as
2788 a bit map, and the map is inverted for negative classes. Previously, a
2789 character > 255 always failed to match such a class; however it should
2790 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2792 (ii) A negated character class with a single character < 255 is coded as
2793 "not this character" (OP_NOT). This wasn't working properly when the test
2794 character was multibyte, either singly or repeated.
2796 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2797 mode, for example: \x{100}{2,3}.
2799 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2800 singly or repeated) now correctly test multibyte characters. However,
2801 PCRE doesn't recognize any characters with values greater than 255 as
2802 digits, spaces, or word characters. Such characters always match \D, \S,
2803 and \W, and never match \d, \s, or \w.
2805 (v) Classes may now contain characters and character ranges with values
2806 greater than 255. For example: [ab\x{100}-\x{400}].
2808 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2809 PCRE in UTF-8 mode.
2811 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2812 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2813 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2814 value.)
2816 53. The single man page has become too large. I have therefore split it up into
2817 a number of separate man pages. These also give rise to individual HTML pages;
2818 these are now put in a separate directory, and there is an index.html page that
2819 lists them all. Some hyperlinking between the pages has been installed.
2821 54. Added convenience functions for handling named capturing parentheses.
2823 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2824 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2825 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2826 are faulted.
2828 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2829 calling configure. These values are used when compiling the dftables.c program
2830 which is run to generate the source of the default character tables. They
2831 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2832 you will need to set these values.
2834 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2837 Version 3.9 02-Jan-02
2838 ---------------------
2840 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2842 2. If --disable-static was given, the building process failed when trying to
2843 build pcretest and pcregrep. (For some reason it was using libtool to compile
2844 them, which is not right, as they aren't part of the library.)
2847 Version 3.8 18-Dec-01
2848 ---------------------
2850 1. The experimental UTF-8 code was completely screwed up. It was packing the
2851 bytes in the wrong order. How dumb can you get?
2854 Version 3.7 29-Oct-01
2855 ---------------------
2857 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2858 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2859 this didn't happen under Solaris 8, where I normally test things.
2861 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2862 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2863 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2864 this fix an hour or so after the initial 3.7 release.)
2867 Version 3.6 23-Oct-01
2868 ---------------------
2870 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2871 offsets passed as NULL with zero offset count.
2873 2. The config.guess and config.sub files had not been updated when I moved to
2874 the latest autoconf.
2877 Version 3.5 15-Aug-01
2878 ---------------------
2880 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2881 had been forgotten.
2883 2. By using declared but undefined structures, we can avoid using "void"
2884 definitions in pcre.h while keeping the internal definitions of the structures
2885 private.
2887 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2888 user point of view, this means that both static and shared libraries are built
2889 by default, but this can be individually controlled. More of the work of
2890 handling this static/shared cases is now inside libtool instead of PCRE's make
2891 file.
2893 4. The pcretest utility is now installed along with pcregrep because it is
2894 useful for users (to test regexs) and by doing this, it automatically gets
2895 relinked by libtool. The documentation has been turned into a man page, so
2896 there are now .1, .txt, and .html versions in /doc.
2898 5. Upgrades to pcregrep:
2899 (i) Added long-form option names like gnu grep.
2900 (ii) Added --help to list all options with an explanatory phrase.
2901 (iii) Added -r, --recursive to recurse into sub-directories.
2902 (iv) Added -f, --file to read patterns from a file.
2904 6. pcre_exec() was referring to its "code" argument before testing that
2905 argument for NULL (and giving an error if it was NULL).
2907 7. Upgraded Makefile.in to allow for compiling in a different directory from
2908 the source directory.
2910 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2911 options bits, the pointer it was passed was to an int instead of to an unsigned
2912 long int. This mattered only on 64-bit systems.
2914 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2915 generated) instead of pcre.in, which it its source. Also made the same change
2916 in several of the .c files.
2918 10. A new release of gcc defines printf() as a macro, which broke pcretest
2919 because it had an ifdef in the middle of a string argument for printf(). Fixed
2920 by using separate calls to printf().
2922 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2923 script, to force use of CR or LF instead of \n in the source. On non-Unix
2924 systems, the value can be set in config.h.
2926 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2927 absolute limit. Changed the text of the error message to make this clear, and
2928 likewise updated the man page.
2930 13. The limit of 99 on the number of capturing subpatterns has been removed.
2931 The new limit is 65535, which I hope will not be a "real" limit.
2934 Version 3.4 22-Aug-00
2935 ---------------------
2937 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2939 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2942 Version 3.3 01-Aug-00
2943 ---------------------
2945 1. If an octal character was given, but the value was greater than \377, it
2946 was not getting masked to the least significant bits, as documented. This could
2947 lead to crashes in some systems.
2949 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2950 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2952 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2953 These just pass their arguments on to (pcre_free)(), but they are provided
2954 because some uses of PCRE bind it to non-C systems that can call its functions,
2955 but cannot call free() or pcre_free() directly.
2957 4. Add "make test" as a synonym for "make check". Corrected some comments in
2958 the Makefile.
2960 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2961 Makefile.
2963 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2964 command called pgrep for grepping around the active processes.
2966 7. Added the beginnings of support for UTF-8 character strings.
2968 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2969 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2970 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2971 out for the ar command.)
2974 Version 3.2 12-May-00
2975 ---------------------
2977 This is purely a bug fixing release.
2979 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2980 of ZA. This was just one example of several cases that could provoke this bug,
2981 which was introduced by change 9 of version 2.00. The code for breaking
2982 infinite loops after an iteration that matches an empty string was't working
2983 correctly.
2985 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2986 when matched against abbab (for example). After matching an empty string, it
2987 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2988 caused it to match further down the string than it should.
2990 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2991 was there because it doesn't seem to be needed, and it causes trouble on some
2992 systems, as it is not a Standard C header. It has been removed.
2994 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2995 were reported on the Macintosh. The changes were from
2997 while ((c = *(++ptr)) != 0 && c != '\n');
2998 to
2999 while ((c = *(++ptr)) != 0 && c != '\n') ;
3001 Totally extraordinary, but if that's what it takes...
3003 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3004 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3005 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3006 assumes the way PCRE uses memmove() (always moving upwards).
3008 6. PCRE is being used in one environment where strchr() is not available. There
3009 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3010 faster code anyway.
3013 Version 3.1 09-Feb-00
3014 ---------------------
3016 The only change in this release is the fixing of some bugs in Makefile.in for
3017 the "install" target:
3019 (1) It was failing to install pcreposix.h.
3021 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3024 Version 3.0 01-Feb-00
3025 ---------------------
3027 1. Add support for the /+ modifier to perltest (to output $` like it does in
3028 pcretest).
3030 2. Add support for the /g modifier to perltest.
3032 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3033 matches null strings.
3035 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3036 pattern. Perl treats empty patterns specially - it reuses the most recent
3037 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3038 effect.
3040 5. The POSIX interface was broken in that it was just handing over the POSIX
3041 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3042 required a bigger vector, with some working space on the end. This means that
3043 the POSIX wrapper now has to get and free some memory, and copy the results.
3045 6. Added some simple autoconf support, placing the test data and the
3046 documentation in separate directories, re-organizing some of the
3047 information files, and making it build pcre-config (a GNU standard). Also added
3048 libtool support for building PCRE as a shared library, which is now the
3049 default.
3051 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3052 09 are not valid octal constants. Single digits will be used for minor values
3053 less than 10.
3055 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3056 existing programs that set these in the POSIX interface can use PCRE without
3057 modification.
3059 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3060 return all that pcre_info() returns, plus additional data. The pcre_info()
3061 function is retained for compatibility, but is considered to be obsolete.
3063 10. Added experimental recursion feature (?R) to handle one common case that
3064 Perl 5.6 will be able to do with (?p{...}).
3066 11. Added support for POSIX character classes like [:alpha:], which Perl is
3067 adopting.
3070 Version 2.08 31-Aug-99
3071 ----------------------
3073 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3074 trying to match at the startoffset position, but instead was moving forward to
3075 the next newline as if a previous match had failed.
3077 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3078 and could get into a loop if a null string was matched other than at the start
3079 of the subject.
3081 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3082 be distinguished at compile time, and for completeness also added PCRE_DATE.
3084 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3085 in GnuWin32 environments.
3088 Version 2.07 29-Jul-99
3089 ----------------------
3091 1. The documentation is now supplied in plain text form and HTML as well as in
3092 the form of man page sources.
3094 2. C++ compilers don't like assigning (void *) values to other pointer types.
3095 In particular this affects malloc(). Although there is no problem in Standard
3096 C, I've put in casts to keep C++ compilers happy.
3098 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3099 should be (const char *).
3101 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3102 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3103 However, I haven't made this a standard facility. The documentation doesn't
3104 mention it, and the Makefile doesn't support it.
3106 5. The Makefile now contains an "install" target, with editable destinations at
3107 the top of the file. The pcretest program is not installed.
3109 6. pgrep -V now gives the PCRE version number and date.
3111 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3112 causing the entire string to be ignored, instead of just the last character.
3114 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3115 non-matching string, it can take a very, very long time, even for strings of
3116 quite modest length, because of the nested recursion. PCRE now does better in
3117 some of these cases. It does this by remembering the last required literal
3118 character in the pattern, and pre-searching the subject to ensure it is present
3119 before running the real match. In other words, it applies a heuristic to detect
3120 some types of certain failure quickly, and in the above example, if presented
3121 with a string that has no trailing " it gives "no match" very quickly.
3123 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3124 other alternatives are tried instead.
3127 Version 2.06 09-Jun-99
3128 ----------------------
3130 1. Change pcretest's output for amount of store used to show just the code
3131 space, because the remainder (the data block) varies in size between 32-bit and
3132 64-bit systems.
3134 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3135 start matching at. This allows lookbehinds to work when searching for multiple
3136 occurrences in a string.
3138 3. Added additional options to pcretest for testing multiple occurrences:
3140 /+ outputs the rest of the string that follows a match
3141 /g loops for multiple occurrences, using the new startoffset argument
3142 /G loops for multiple occurrences by passing an incremented pointer
3144 4. PCRE wasn't doing the "first character" optimization for patterns starting
3145 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3146 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3147 the letter 'x'. On long subject strings, this gives a significant speed-up.
3150 Version 2.05 21-Apr-99
3151 ----------------------
3153 1. Changed the type of magic_number from int to long int so that it works
3154 properly on 16-bit systems.
3156 2. Fixed a bug which caused patterns starting with .* not to work correctly
3157 when the subject string contained newline characters. PCRE was assuming
3158 anchoring for such patterns in all cases, which is not correct because .* will
3159 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3160 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3161 must be retried after every newline in the subject.
3164 Version 2.04 18-Feb-99
3165 ----------------------
3167 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3168 computation of the store needed to hold the pattern was incorrect (too large).
3169 If such patterns were nested a few deep, this could multiply and become a real
3170 problem.
3172 2. Added /M option to pcretest to show the memory requirement of a specific
3173 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3175 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3176 compiled in such a way that the backtracking after subsequent failure was
3177 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3178 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3181 Version 2.03 02-Feb-99
3182 ----------------------
3184 1. Fixed typo and small mistake in man page.
3186 2. Added 4th condition (GPL supersedes if conflict) and created separate
3187 LICENCE file containing the conditions.
3189 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3190 Perl, that is the internal \ allows the delimiter to be included in the
3191 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3192 the final delimiter, add \ to the end of the pattern (to test the error).
3194 4. Added the convenience functions for extracting substrings after a successful
3195 match. Updated pcretest to make it able to test these functions.
3198 Version 2.02 14-Jan-99
3199 ----------------------
3201 1. Initialized the working variables associated with each extraction so that
3202 their saving and restoring doesn't refer to uninitialized store.
3204 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3205 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3206 fix the problem.
3208 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3209 calls, and wasn't printing the correct value for compiling calls. Increased the
3210 default value of LOOPREPEAT, and the number of significant figures in the
3211 times.
3213 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3215 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3216 a building problem on Windows NT with a FAT file system.
3219 Version 2.01 21-Oct-98
3220 ----------------------
3222 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3223 to character tables built by pcre_maketables() in the current locale. If NULL
3224 is passed, the default tables are used.
3227 Version 2.00 24-Sep-98
3228 ----------------------
3230 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3231 it any more.
3233 2. Allow quantification of (?>) groups, and make it work correctly.
3235 3. The first character computation wasn't working for (?>) groups.
3237 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3238 end of the subject) and add 5.005's \z, which really does match only at the
3239 very end of the subject.
3241 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3243 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3244 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3245 localized options. All options to pcre_study() were also removed.
3247 7. Add other new features from 5.005:
3249 $(?<= positive lookbehind
3250 $(?<! negative lookbehind
3251 (?imsx-imsx) added the unsetting capability
3252 such a setting is global if at outer level; local otherwise
3253 (?imsx-imsx:) non-capturing groups with option setting
3254 (?(cond)re|re) conditional pattern matching
3256 A backreference to itself in a repeated group matches the previous
3257 captured string.
3259 8. General tidying up of studying (both automatic and via "study")
3260 consequential on the addition of new assertions.
3262 9. As in 5.005, unlimited repeated groups that could match an empty substring
3263 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3264 runtime if any iteration does actually match an empty substring.
3266 10. Include the RunTest script in the distribution.
3268 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3269 discrepancies, some of which were old and were also with respect to 5.004. They
3270 have now been fixed.
3273 Version 1.09 28-Apr-98
3274 ----------------------
3276 1. A negated single character class followed by a quantifier with a minimum
3277 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3278 program crashes, or just wrong answers. This did not apply to negated classes
3279 containing more than one character, or to minima other than one.
3282 Version 1.08 27-Mar-98
3283 ----------------------
3285 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3287 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3288 latter must appear before anything that relies on it in the pattern.
3291 Version 1.07 16-Feb-98
3292 ----------------------
3294 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3295 repeat of a potentially empty string).
3298 Version 1.06 23-Jan-98
3299 ----------------------
3301 1. Added Markus Oberhumer's little patches for C++.
3303 2. Literal strings longer than 255 characters were broken.
3306 Version 1.05 23-Dec-97
3307 ----------------------
3309 1. Negated character classes containing more than one character were failing if
3310 PCRE_CASELESS was set at run time.
3313 Version 1.04 19-Dec-97
3314 ----------------------
3316 1. Corrected the man page, where some "const" qualifiers had been omitted.
3318 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3319 input syntax.
3321 3. Fixed memory leak which occurred when a regex with back references was
3322 matched with an offsets vector that wasn't big enough. The temporary memory
3323 that is used in this case wasn't being freed if the match failed.
3325 4. Tidied pcretest to ensure it frees memory that it gets.
3327 5. Temporary memory was being obtained in the case where the passed offsets
3328 vector was exactly big enough.
3330 6. Corrected definition of offsetof() from change 5 below.
3332 7. I had screwed up change 6 below and broken the rules for the use of
3333 setjmp(). Now fixed.
3336 Version 1.03 18-Dec-97
3337 ----------------------
3339 1. A erroneous regex with a missing opening parenthesis was correctly
3340 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3341 on some systems.
3343 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3344 it was reported that one broken compiler failed on the former because "code" is
3345 also an independent variable.
3347 3. The erroneous regex a[]b caused an array overrun reference.
3349 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3350 fail on data ending with that character. (It was going on too far, and checking
3351 the next character, typically a binary zero.) This was specific to the
3352 optimized code for single-character negative classes.
3354 5. Added a contributed patch from the TIN world which does the following:
3356 + Add an undef for memmove, in case the the system defines a macro for it.
3358 + Add a definition of offsetof(), in case there isn't one. (I don't know
3359 the reason behind this - offsetof() is part of the ANSI standard - but
3360 it does no harm).
3362 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3363 most of the places where whitespace preceded '#'. I have given up and
3364 allowed the remaining 2 cases to be at the margin.
3366 + Rename some variables in pcre to eliminate shadowing. This seems very
3367 pedantic, but does no harm, of course.
3369 6. Moved the call to setjmp() into its own function, to get rid of warnings
3370 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3372 7. Constructs such as \d{8,} were compiling into the equivalent of
3373 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3374 outcome, but in this particular case used more store than had been allocated,
3375 which caused the bug to be discovered because it threw up an internal error.
3377 8. The debugging code in both pcre and pcretest for outputting the compiled
3378 form of a regex was going wrong in the case of back references followed by
3379 curly-bracketed repeats.
3382 Version 1.02 12-Dec-97
3383 ----------------------
3385 1. Typos in pcre.3 and comments in the source fixed.
3387 2. Applied a contributed patch to get rid of places where it used to remove
3388 'const' from variables, and fixed some signed/unsigned and uninitialized
3389 variable warnings.
3391 3. Added the "runtest" target to Makefile.
3393 4. Set default compiler flag to -O2 rather than just -O.
3396 Version 1.01 19-Nov-97
3397 ----------------------
3399 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3400 like /([ab]*)*/, that is, for classes with more than one character in them.
3402 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3403 as /((?>a*))*/ (a PCRE_EXTRA facility).
3406 Version 1.00 18-Nov-97
3407 ----------------------
3409 1. Added compile-time macros to support systems such as SunOS4 which don't have
3410 memmove() or strerror() but have other things that can be used instead.
3412 2. Arranged that "make clean" removes the executables.
3415 Version 0.99 27-Oct-97
3416 ----------------------
3418 1. Fixed bug in code for optimizing classes with only one character. It was
3419 initializing a 32-byte map regardless, which could cause it to run off the end
3420 of the memory it had got.
3422 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3425 Version 0.98 22-Oct-97
3426 ----------------------
3428 1. Fixed bug in code for handling temporary memory usage when there are more
3429 back references than supplied space in the ovector. This could cause segfaults.
3432 Version 0.97 21-Oct-97
3433 ----------------------
3435 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3437 2. Optimized negated single characters not to use a bit map.
3439 3. Brought error texts together as macro definitions; clarified some of them;
3440 fixed one that was wrong - it said "range out of order" when it meant "invalid
3441 escape sequence".
3443 4. Changed some char * arguments to const char *.
3445 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3447 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3448 pcretest.
3451 Version 0.96 16-Oct-97
3452 ----------------------
3454 1. Added a simple "pgrep" utility to the distribution.
3456 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3457 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3458 where "ddd" means "one or more decimal digits".
3460 3. Fixed serious bug. If a pattern had a back reference, but the call to
3461 pcre_exec() didn't supply a large enough ovector to record the related
3462 identifying subpattern, the match always failed. PCRE now remembers the number
3463 of the largest back reference, and gets some temporary memory in which to save
3464 the offsets during matching if necessary, in order to ensure that
3465 backreferences always work.
3467 4. Increased the compatibility with Perl in a number of ways:
3469 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3470 to request this handling. The option can be set at compile or exec time.
3472 (b) $ matches before a terminating newline by default; an option
3473 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3474 mode). The option can be set at compile or exec time.
3476 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3477 the same as Perl's. If the decimal number it represents is less than 10
3478 or there aren't that many previous left capturing parentheses, an octal
3479 escape is read. Inside a character class, it's always an octal escape,
3480 even if it is a single digit.
3482 (d) An escaped but undefined alphabetic character is taken as a literal,
3483 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3484 escapes.
3486 (e) {0} is now permitted. (The previous item is removed from the compiled
3487 pattern).
3489 5. Changed all the names of code files so that the basic parts are no longer
3490 than 10 characters, and abolished the teeny "globals.c" file.
3492 6. Changed the handling of character classes; they are now done with a 32-byte
3493 bit map always.
3495 7. Added the -d and /D options to pcretest to make it possible to look at the
3496 internals of compilation without having to recompile pcre.
3499 Version 0.95 23-Sep-97
3500 ----------------------
3502 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3503 \x20 at the start of a run of normal characters. These were being treated as
3504 real characters, instead of the source characters being re-checked.
3507 Version 0.94 18-Sep-97
3508 ----------------------
3510 1. The functions are now thread-safe, with the caveat that the global variables
3511 containing pointers to malloc() and free() or alternative functions are the
3512 same for all threads.
3514 2. Get pcre_study() to generate a bitmap of initial characters for non-
3515 anchored patterns when this is possible, and use it if passed to pcre_exec().
3518 Version 0.93 15-Sep-97
3519 ----------------------
3521 1. /(b)|(:+)/ was computing an incorrect first character.
3523 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3524 but not actually doing anything yet.
3526 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3527 as Perl does (e.g. [-az] or [az-]).
3529 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3530 all possible positions.
3532 5. Split up into different modules to avoid including unneeded functions in a
3533 compiled binary. However, compile and exec are still in one module. The "study"
3534 function is split off.
3536 6. The character tables are now in a separate module whose source is generated
3537 by an auxiliary program - but can then be edited by hand if required. There are
3538 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3539 toupper() in the code.
3541 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3542 make them global. Abolish the function for setting them, as the caller can now
3543 set them directly.
3546 Version 0.92 11-Sep-97
3547 ----------------------
3549 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3550 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3552 2. Caseless matching was not working in character classes if the characters in
3553 the pattern were in upper case.
3555 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3557 4. Make PCRE_ANCHORED public and accept as a compile option.
3559 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3560 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3561 pass them.
3563 6. Give an error if bad option bits passed at compile or run time.
3565 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3566 pcretest to cause it to pass that flag.
3568 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3569 options, and the first character, if set.
3571 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3574 Version 0.91 10-Sep-97
3575 ----------------------
3577 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3578 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3580 2. PCRE was looping on encountering an indefinitely repeated back reference to
3581 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3582 Perl does - treats the match as successful.
3584 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5