ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 351 - (show annotations)
Fri Jul 4 18:27:16 2008 UTC (12 years, 7 months ago) by ph10
File size: 156670 byte(s)
Error occurred while calculating annotation data.
Final tidies for new Unicode property code; upgrade to Unicode 5.1.0.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.0 02 Jul-08
5 ---------------------
7 1. Replaced UCP searching code with optimized version as implemented for Ad
8 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
9 stage table and inline lookup instead of a function, giving speed ups of 2
10 to 5 times on some simple patterns that I tested. Permission was given to
11 distribute the MultiStage2.py script that generates the tables (it's not in
12 the tarball, but is in the Subversion repository).
14 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
15 scripts.
18 Version 7.7 07-May-08
19 ---------------------
21 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
22 a string to a long long, pretend we don't even have a long long." This is
23 done by checking for the strtoq, strtoll, and _strtoi64 functions.
25 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
26 pre-7.6 versions, which defined a global no_arg variable instead of putting
27 it in the RE class. (See also #8 below.)
29 3. Remove a line of dead code, identified by coverity and reported by Nuno
30 Lopes.
32 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
34 (1) The include/exclude patterns were being applied to the whole pathnames
35 of files, instead of just to the final components.
37 (2) If there was more than one level of directory, the subdirectories were
38 skipped unless they satisfied the include/exclude conditions. This is
39 inconsistent with GNU grep (and could even be seen as contrary to the
40 pcregrep specification - which I improved to make it absolutely clear).
41 The action now is always to scan all levels of directory, and just
42 apply the include/exclude patterns to regular files.
44 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
45 --exclude_dir in the tests to avoid scanning .svn directories.
47 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
48 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
49 doesn't support NULs in patterns.
51 7. Added some missing "const"s to declarations of static tables in
52 pcre_compile.c and pcre_dfa_exec.c.
54 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
55 caused by fix #2 above. (Subsequently also a second patch to fix the
56 first patch. And a third patch - this was a messy problem.)
58 9. Applied Craig's patch to remove the use of push_back().
60 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
61 matching function regexec().
63 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
64 which, however, unlike Perl's \g{...}, are subroutine calls, not back
65 references. PCRE supports relative numbers with this syntax (I don't think
66 Oniguruma does).
68 12. Previously, a group with a zero repeat such as (...){0} was completely
69 omitted from the compiled regex. However, this means that if the group
70 was called as a subroutine from elsewhere in the pattern, things went wrong
71 (an internal error was given). Such groups are now left in the compiled
72 pattern, with a new opcode that causes them to be skipped at execution
73 time.
75 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
76 to the way PCRE behaves:
78 (a) A lone ] character is dis-allowed (Perl treats it as data).
80 (b) A back reference to an unmatched subpattern matches an empty string
81 (Perl fails the current match path).
83 (c) A data ] in a character class must be notated as \] because if the
84 first data character in a class is ], it defines an empty class. (In
85 Perl it is not possible to have an empty class.) The empty class []
86 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
87 The negative empty class [^] matches any one character, independently
88 of the DOTALL setting.
90 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
91 non-existent subpattern following a character class starting with ']' and
92 containing () gave an internal compiling error instead of "reference to
93 non-existent subpattern". Fortunately, when the pattern did exist, the
94 compiled code was correct. (When scanning forwards to check for the
95 existencd of the subpattern, it was treating the data ']' as terminating
96 the class, so got the count wrong. When actually compiling, the reference
97 was subsequently set up correctly.)
99 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
100 it was being rejected as not supported by pcre_dfa_exec(), even though
101 other assertions are supported. I have made pcre_dfa_exec() support
102 (*FAIL).
104 16. The implementation of 13c above involved the invention of a new opcode,
105 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
106 cannot be changed at match time, I realized I could make a small
107 improvement to matching performance by compiling OP_ALLANY instead of
108 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
109 on the OP_ANY path.
111 17. Compiling pcretest on Windows with readline support failed without the
112 following two fixes: (1) Make the unistd.h include conditional on
113 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
115 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
116 ncurses library to be included for pcretest when ReadLine support is
117 requested, but also to allow for it to be overridden. This patch came from
118 Daniel Bergström.
120 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
121 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
122 any errors with the current Unicode tables. Thanks to Peter Kankowski for
123 spotting this.
126 Version 7.6 28-Jan-08
127 ---------------------
129 1. A character class containing a very large number of characters with
130 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
131 overflow.
133 2. Patch to cut out the "long long" test in pcrecpp_unittest when
134 HAVE_LONG_LONG is not defined.
136 3. Applied Christian Ehrlicher's patch to update the CMake build files to
137 bring them up to date and include new features. This patch includes:
139 - Fixed PH's badly added libz and libbz2 support.
140 - Fixed a problem with static linking.
141 - Added pcredemo. [But later removed - see 7 below.]
142 - Fixed dftables problem and added an option.
143 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
145 - Added readline support for pcretest.
146 - Added an listing of the option settings after cmake has run.
148 4. A user submitted a patch to Makefile that makes it easy to create
149 "pcre.dll" under mingw when using Configure/Make. I added stuff to
150 Makefile.am that cause it to include this special target, without
151 affecting anything else. Note that the same mingw target plus all
152 the other distribution libraries and programs are now supported
153 when configuring with CMake (see 6 below) instead of with
154 Configure/Make.
156 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
157 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
158 exported in the Windows port". It has not yet been confirmed that the patch
159 solves the problem, but it does no harm.
161 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
162 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
163 with CMake, and also correct the comment about stack recursion.
165 7. Remove the automatic building of pcredemo from the ./configure system and
166 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
167 of a program that users should build themselves after PCRE is installed, so
168 building it automatically is not really right. What is more, it gave
169 trouble in some build environments.
171 8. Further tidies to CMakeLists.txt from Sheri and Christian.
174 Version 7.5 10-Jan-08
175 ---------------------
177 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
178 values in parens when parsing an RE using the C++ wrapper."
180 2. Negative specials like \S did not work in character classes in UTF-8 mode.
181 Characters greater than 255 were excluded from the class instead of being
182 included.
184 3. The same bug as (2) above applied to negated POSIX classes such as
185 [:^space:].
187 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
188 defined or documented. It seems to have been a typo for PCRE_STATIC, so
189 I have changed it.
191 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
192 first named subpattern) and a construct such as (?&a) would reference the
193 first named subpattern whose name started with "a" (in other words, the
194 length check was missing). Both these problems are fixed. "Subpattern name
195 expected" is now given for (?&) (a zero-length name), and this patch also
196 makes it give the same error for \k'' (previously it complained that that
197 was a reference to a non-existent subpattern).
199 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
200 this is right because (?- can be followed by option settings as well as by
201 digits. I have, however, made the messages clearer.
203 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
204 than the number used in the conditional) now cause a compile-time error.
205 This is actually not compatible with Perl, which accepts such patterns, but
206 treats the conditional as always being FALSE (as PCRE used to), but it
207 seems to me that giving a diagnostic is better.
209 8. Change "alphameric" to the more common word "alphanumeric" in comments
210 and messages.
212 9. Fix two occurrences of "backslash" in comments that should have been
213 "backspace".
215 10. Remove two redundant lines of code that can never be obeyed (their function
216 was moved elsewhere).
218 11. The program that makes PCRE's Unicode character property table had a bug
219 which caused it to generate incorrect table entries for sequences of
220 characters that have the same character type, but are in different scripts.
221 It amalgamated them into a single range, with the script of the first of
222 them. In other words, some characters were in the wrong script. There were
223 thirteen such cases, affecting characters in the following ranges:
225 U+002b0 - U+002c1
226 U+0060c - U+0060d
227 U+0061e - U+00612
228 U+0064b - U+0065e
229 U+0074d - U+0076d
230 U+01800 - U+01805
231 U+01d00 - U+01d77
232 U+01d9b - U+01dbf
233 U+0200b - U+0200f
234 U+030fc - U+030fe
235 U+03260 - U+0327f
236 U+0fb46 - U+0fbb1
237 U+10450 - U+1049d
239 12. The -o option (show only the matching part of a line) for pcregrep was not
240 compatible with GNU grep in that, if there was more than one match in a
241 line, it showed only the first of them. It now behaves in the same way as
242 GNU grep.
244 13. If the -o and -v options were combined for pcregrep, it printed a blank
245 line for every non-matching line. GNU grep prints nothing, and pcregrep now
246 does the same. The return code can be used to tell if there were any
247 non-matching lines.
249 14. Added --file-offsets and --line-offsets to pcregrep.
251 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
252 infinitely looping recursion. The bug was that positive lookaheads were not
253 being skipped when checking for a possible empty match (negative lookaheads
254 and both kinds of lookbehind were skipped).
256 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
257 inclusion of <windows.h> to before rather than after the definition of
258 INVALID_FILE_ATTRIBUTES (patch from David Byron).
260 17. Specifying a possessive quantifier with a specific limit for a Unicode
261 character property caused pcre_compile() to compile bad code, which led at
262 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
263 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
264 caused the error; without that there was no problem.
266 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
268 19. Added --enable-pcretest-libreadline.
270 20. In pcrecpp.cc, the variable 'count' was incremented twice in
271 RE::GlobalReplace(). As a result, the number of replacements returned was
272 double what it should be. I removed one of the increments, but Craig sent a
273 later patch that removed the other one (the right fix) and added unit tests
274 that check the return values (which was not done before).
276 21. Several CMake things:
278 (1) Arranged that, when cmake is used on Unix, the libraries end up with
279 the names libpcre and libpcreposix, not just pcre and pcreposix.
281 (2) The above change means that pcretest and pcregrep are now correctly
282 linked with the newly-built libraries, not previously installed ones.
286 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
287 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
288 UTF-8 newline character). The key issue is that the pattern starts .*;
289 this means that the match must be either at the beginning, or after a
290 newline. The bug was in the code for advancing after a failed match and
291 checking that the new position followed a newline. It was not taking
292 account of UTF-8 characters correctly.
294 23. PCRE was behaving differently from Perl in the way it recognized POSIX
295 character classes. PCRE was not treating the sequence [:...:] as a
296 character class unless the ... were all letters. Perl, however, seems to
297 allow any characters between [: and :], though of course it rejects as
298 unknown any "names" that contain non-letters, because all the known class
299 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
300 for example, whereas PCRE did not - it did not recognize a POSIX character
301 class. This seemed a bit dangerous, so the code has been changed to be
302 closer to Perl. The behaviour is not identical to Perl, because PCRE will
303 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
304 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
305 Perl does, and where it didn't before.
307 24. Rewrite so as to remove the single use of %n from pcregrep because in some
308 Windows environments %n is disabled by default.
311 Version 7.4 21-Sep-07
312 ---------------------
314 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
315 means that a class such as [\s] counted as "explicit reference to CR or
316 LF". That isn't really right - the whole point of the change was to try to
317 help when there was an actual mention of one of the two characters. So now
318 the change happens only if \r or \n (or a literal CR or LF) character is
319 encountered.
321 2. The 32-bit options word was also used for 6 internal flags, but the numbers
322 of both had grown to the point where there were only 3 bits left.
323 Fortunately, there was spare space in the data structure, and so I have
324 moved the internal flags into a new 16-bit field to free up more option
325 bits.
327 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
328 but did not set the internal JCHANGED flag - either of these is enough to
329 control the way the "get" function works - but the PCRE_INFO_JCHANGED
330 facility is supposed to tell if (?J) was ever used, so now (?J) at the
331 start sets both bits.
333 4. Added options (at build time, compile time, exec time) to change \R from
334 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
336 5. doc/pcresyntax.html was missing from the distribution.
338 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
339 compatibility, even though it is no longer used.
341 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
342 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
343 windows.h file is present (where different names are used). [This was
344 reversed later after testing - see 16 below.]
346 8. Changed all #include <config.h> to #include "config.h". There were also
347 some further <pcre.h> cases that I changed to "pcre.h".
349 9. When pcregrep was used with the --colour option, it missed the line ending
350 sequence off the lines that it output.
352 10. It was pointed out to me that arrays of string pointers cause lots of
353 relocations when a shared library is dynamically loaded. A technique of
354 using a single long string with a table of offsets can drastically reduce
355 these. I have refactored PCRE in four places to do this. The result is
356 dramatic:
358 Originally: 290
359 After changing UCP table: 187
360 After changing error message table: 43
361 After changing table of "verbs" 36
362 After changing table of Posix names 22
364 Thanks to the folks working on Gregex for glib for this insight.
366 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
367 unicode-properties was also set.
369 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
371 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
372 checked only for CRLF.
374 14. Added casts to pcretest.c to avoid compiler warnings.
376 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
378 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
379 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
380 entirely. This removes changes made in 7 above.
382 17. The CMake files have been updated, and there is now more information about
383 building with CMake in the NON-UNIX-USE document.
386 Version 7.3 28-Aug-07
387 ---------------------
389 1. In the rejigging of the build system that eventually resulted in 7.1, the
390 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
391 brackets there is not right, since it causes compilers to look for an
392 installed pcre.h, not the version that is in the source that is being
393 compiled (which of course may be different). I have changed it back to:
395 #include "pcre.h"
397 I have a vague recollection that the change was concerned with compiling in
398 different directories, but in the new build system, that is taken care of
399 by the VPATH setting the Makefile.
401 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
402 when the subject happened to end in the byte 0x85 (e.g. if the last
403 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
404 characters but of course it shouldn't be taken as a newline when it is part
405 of another character. The bug was that, for an unlimited repeat of . in
406 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
407 characters when looking for a newline.
409 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
411 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
412 in debug output.
414 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
415 long printing in the pcrecpp unittest when running under MinGW.
417 6. ESC_K was left out of the EBCDIC table.
419 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
420 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
421 limit also applies to "virtual nesting" when a pattern is recursive, and in
422 this case 1000 isn't so big. I have been able to remove this limit at the
423 expense of backing off one optimization in certain circumstances. Normally,
424 when pcre_exec() would call its internal match() function recursively and
425 immediately return the result unconditionally, it uses a "tail recursion"
426 feature to save stack. However, when a subpattern that can match an empty
427 string has an unlimited repetition quantifier, it no longer makes this
428 optimization. That gives it a stack frame in which to save the data for
429 checking that an empty string has been matched. Previously this was taken
430 from the 1000-entry workspace that had been reserved. So now there is no
431 explicit limit, but more stack is used.
433 8. Applied Daniel's patches to solve problems with the import/export magic
434 syntax that is required for Windows, and which was going wrong for the
435 pcreposix and pcrecpp parts of the library. These were overlooked when this
436 problem was solved for the main library.
438 9. There were some crude static tests to avoid integer overflow when computing
439 the size of patterns that contain repeated groups with explicit upper
440 limits. As the maximum quantifier is 65535, the maximum group length was
441 set at 30,000 so that the product of these two numbers did not overflow a
442 32-bit integer. However, it turns out that people want to use groups that
443 are longer than 30,000 bytes (though not repeat them that many times).
444 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
445 made it possible to implement the integer overflow checks in a much more
446 dynamic way, which I have now done. The artificial limitation on group
447 length has been removed - we now have only the limit on the total length of
448 the compiled pattern, which depends on the LINK_SIZE setting.
450 10. Fixed a bug in the documentation for get/copy named substring when
451 duplicate names are permitted. If none of the named substrings are set, the
452 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
453 empty string.
455 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
456 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
457 because the ] is interpreted as the first data character and the
458 terminating ] is not found. PCRE has been made compatible with Perl in this
459 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
460 cause memory overwriting.
462 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
463 string has been matched (to stop an infinite loop). It was not recognizing
464 a conditional subpattern that could match an empty string if that
465 subpattern was within another subpattern. For example, it looped when
466 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
467 condition was not nested. This bug has been fixed.
469 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
470 past the start of the subject in the presence of bytes with the top bit
471 set, for example "\x8aBCD".
473 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
474 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
476 14. Optimized (?!) to (*FAIL).
478 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
479 This restricts code points to be within the range 0 to 0x10FFFF, excluding
480 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
481 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
482 does: it's just the validity check that is more restrictive.
484 16. Inserted checks for integer overflows during escape sequence (backslash)
485 processing, and also fixed erroneous offset values for syntax errors during
486 backslash processing.
488 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
489 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
491 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
492 caused an overrun.
494 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
495 something other than just ASCII characters) inside a group that had an
496 unlimited repeat caused a loop at compile time (while checking to see
497 whether the group could match an empty string).
499 20. Debugging a pattern containing \p or \P could cause a crash. For example,
500 [\P{Any}] did so. (Error in the code for printing property names.)
502 21. An orphan \E inside a character class could cause a crash.
504 22. A repeated capturing bracket such as (A)? could cause a wild memory
505 reference during compilation.
507 23. There are several functions in pcre_compile() that scan along a compiled
508 expression for various reasons (e.g. to see if it's fixed length for look
509 behind). There were bugs in these functions when a repeated \p or \P was
510 present in the pattern. These operators have additional parameters compared
511 with \d, etc, and these were not being taken into account when moving along
512 the compiled data. Specifically:
514 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
515 length.
517 (b) An item such as \pL+ within a repeated group could cause crashes or
518 loops.
520 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
521 "reference to non-existent subpattern" error.
523 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
525 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
526 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
528 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
530 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
531 character were causing crashes (broken optimization).
533 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
534 \p or \P) caused a compile-time loop.
536 28. More problems have arisen in unanchored patterns when CRLF is a valid line
537 break. For example, the unstudied pattern [\r\n]A does not match the string
538 "\r\nA" because change 7.0/46 below moves the current point on by two
539 characters after failing to match at the start. However, the pattern \nA
540 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
541 the same is true. There doesn't seem any very clean way out of this, but
542 what I have chosen to do makes the common cases work: PCRE now takes note
543 of whether there can be an explicit match for \r or \n anywhere in the
544 pattern, and if so, 7.0/46 no longer applies. As part of this change,
545 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
546 pattern has explicit CR or LF references.
548 29. Added (*CR) etc for changing newline setting at start of pattern.
551 Version 7.2 19-Jun-07
552 ---------------------
554 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
555 which is apparently normally available under Windows.
557 2. Re-jig the pcregrep tests with different newline settings in an attempt
558 to make them independent of the local environment's newline setting.
560 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
562 4. Some of the "internals" tests were previously cut out when the link size
563 was not 2, because the output contained actual offsets. The recent new
564 "Z" feature of pcretest means that these can be cut out, making the tests
565 usable with all link sizes.
567 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
568 stack recursion. This gives a massive performance boost under BSD, but just
569 a small improvement under Linux. However, it saves one field in the frame
570 in all cases.
572 6. Added more features from the forthcoming Perl 5.10:
574 (a) (?-n) (where n is a string of digits) is a relative subroutine or
575 recursion call. It refers to the nth most recently opened parentheses.
577 (b) (?+n) is also a relative subroutine call; it refers to the nth next
578 to be opened parentheses.
580 (c) Conditions that refer to capturing parentheses can be specified
581 relatively, for example, (?(-2)... or (?(+3)...
583 (d) \K resets the start of the current match so that everything before
584 is not part of it.
586 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
588 (f) \g{name} is another synonym - part of Perl 5.10's unification of
589 reference syntax.
591 (g) (?| introduces a group in which the numbering of parentheses in each
592 alternative starts with the same number.
594 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
596 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
599 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
600 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
601 for detecting groups that can match an empty string.
603 9. A pattern with a very large number of alternatives (more than several
604 hundred) was running out of internal workspace during the pre-compile
605 phase, where pcre_compile() figures out how much memory will be needed. A
606 bit of new cunning has reduced the workspace needed for groups with
607 alternatives. The 1000-alternative test pattern now uses 12 bytes of
608 workspace instead of running out of the 4096 that are available.
610 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
612 11. Applied patch from Google to remove an optimization that didn't quite work.
613 The report of the bug said:
615 pcrecpp::RE("a*").FullMatch("aaa") matches, while
616 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
617 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
619 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
620 it matched the wrong number of bytes.
623 Version 7.1 24-Apr-07
624 ---------------------
626 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
627 that is more "standard", making use of automake and other Autotools. There
628 is some re-arrangement of the files and adjustment of comments consequent
629 on this.
631 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
632 for recursive directory scanning broke on some systems because the files
633 are not scanned in any specific order and on different systems the order
634 was different. A call to "sort" has been inserted into RunGrepTest for the
635 approprate test as a short-term fix. In the longer term there may be an
636 alternative.
638 3. I had an email from Eric Raymond about problems translating some of PCRE's
639 man pages to HTML (despite the fact that I distribute HTML pages, some
640 people do their own conversions for various reasons). The problems
641 concerned the use of low-level troff macros .br and .in. I have therefore
642 removed all such uses from the man pages (some were redundant, some could
643 be replaced by .nf/.fi pairs). The 132html script that I use to generate
644 HTML has been updated to handle .nf/.fi and to complain if it encounters
645 .br or .in.
647 4. Updated comments in configure.ac that get placed in config.h.in and also
648 arranged for config.h to be included in the distribution, with the name
649 config.h.generic, for the benefit of those who have to compile without
650 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
652 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
653 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
654 makevp.bat for the current PCRE, using the additional files
655 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
657 6. A Windows user reported a minor discrepancy with test 2, which turned out
658 to be caused by a trailing space on an input line that had got lost in his
659 copy. The trailing space was an accident, so I've just removed it.
661 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
662 that is needed.
664 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
665 as "const" (a) because they are and (b) because it helps the PHP
666 maintainers who have recently made a script to detect big data structures
667 in the php code that should be moved to the .rodata section. I remembered
668 to update Builducptable as well, so it won't revert if ucptable.h is ever
669 re-created.
671 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
672 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
673 order to be able to cut out the UTF-8 tables in the latter when UTF-8
674 support is not required. This saves 1.5-2K of code, which is important in
675 some applications.
677 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
678 so as not to refer to the tables, even though these functions will never be
679 called when UTF-8 support is disabled. Otherwise there are problems with a
680 shared library.
682 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
684 (a) It was defining its arguments as char * instead of void *.
686 (b) It was assuming that all moves were upwards in memory; this was true
687 a long time ago when I wrote it, but is no longer the case.
689 The emulated memove() is provided for those environments that have neither
690 memmove() nor bcopy(). I didn't think anyone used it these days, but that
691 is clearly not the case, as these two bugs were recently reported.
693 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
694 and Detrail to create the HTML documentation, the .txt form of the man
695 pages, and it removes trailing spaces from listed files. It also creates
696 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
697 case, it wraps all the #defines with #ifndefs. This script should be run
698 before "make dist".
700 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
701 with Unicode property support.
703 (a) For a maximizing quantifier, if the two different cases of the
704 character were of different lengths in their UTF-8 codings (there are
705 some cases like this - I found 11), and the matching function had to
706 back up over a mixture of the two cases, it incorrectly assumed they
707 were both the same length.
709 (b) When PCRE was configured to use the heap rather than the stack for
710 recursion during matching, it was not correctly preserving the data for
711 the other case of a UTF-8 character when checking ahead for a match
712 while processing a minimizing repeat. If the check also involved
713 matching a wide character, but failed, corruption could cause an
714 erroneous result when trying to check for a repeat of the original
715 character.
717 13. Some tidying changes to the testing mechanism:
719 (a) The RunTest script now detects the internal link size and whether there
720 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
721 values substituted by "configure". (The RunGrepTest script already did
722 this for UTF-8.) The configure.ac script no longer substitutes the
723 relevant variables.
725 (b) The debugging options /B and /D in pcretest show the compiled bytecode
726 with length and offset values. This means that the output is different
727 for different internal link sizes. Test 2 is skipped for link sizes
728 other than 2 because of this, bypassing the problem. Unfortunately,
729 there was also a test in test 3 (the locale tests) that used /B and
730 failed for link sizes other than 2. Rather than cut the whole test out,
731 I have added a new /Z option to pcretest that replaces the length and
732 offset values with spaces. This is now used to make test 3 independent
733 of link size. (Test 2 will be tidied up later.)
735 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
736 segmentation fault instead of returning the appropriate error message.
738 15. In multiline mode when the newline sequence was set to "any", the pattern
739 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
740 This doesn't seem right; it now treats the CRLF combination as the line
741 ending, and so does not match in that case. It's only a pattern such as ^$
742 that would hit this one: something like ^ABC$ would have failed after \r
743 and then tried again after \r\n.
745 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
746 in an attempt to make files that differ only in their line terminators
747 compare equal. This works on Linux.
749 17. Under certain error circumstances pcregrep might try to free random memory
750 as it exited. This is now fixed, thanks to valgrind.
752 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
753 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
754 was because its rules for how to advance for /g after matching an empty
755 string at the end of a line did not allow for this case. They now check for
756 it specially.
758 20. pcretest is supposed to handle patterns and data of any length, by
759 extending its buffers when necessary. It was getting this wrong when the
760 buffer for a data line had to be extended.
762 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
763 CRLF as a newline sequence.
765 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
766 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
767 I have nevertheless tidied it up.
769 23. Added some casts to kill warnings from HP-UX ia64 compiler.
771 24. Added a man page for pcre-config.
774 Version 7.0 19-Dec-06
775 ---------------------
777 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
778 moving to gcc 4.1.1.
780 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
781 sys/time.h, which is documented as needed for this function. It doesn't
782 seem to matter on Linux, but it showed up on some releases of OS X.
784 3. It seems that there are systems where bytes whose values are greater than
785 127 match isprint() in the "C" locale. The "C" locale should be the
786 default when a C program starts up. In most systems, only ASCII printing
787 characters match isprint(). This difference caused the output from pcretest
788 to vary, making some of the tests fail. I have changed pcretest so that:
790 (a) When it is outputting text in the compiled version of a pattern, bytes
791 other than 32-126 are always shown as hex escapes.
793 (b) When it is outputting text that is a matched part of a subject string,
794 it does the same, unless a different locale has been set for the match
795 (using the /L modifier). In this case, it uses isprint() to decide.
797 4. Fixed a major bug that caused incorrect computation of the amount of memory
798 required for a compiled pattern when options that changed within the
799 pattern affected the logic of the preliminary scan that determines the
800 length. The relevant options are -x, and -i in UTF-8 mode. The result was
801 that the computed length was too small. The symptoms of this bug were
802 either the PCRE error "internal error: code overflow" from pcre_compile(),
803 or a glibc crash with a message such as "pcretest: free(): invalid next
804 size (fast)". Examples of patterns that provoked this bug (shown in
805 pcretest format) are:
807 /(?-x: )/x
808 /(?x)(?-x: \s*#\s*)/
809 /((?i)[\x{c0}])/8
810 /(?i:[\x{c0}])/8
812 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
813 is now done differently.
815 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
816 wrapper classes; (b) implement a new function in the C++ scanner that is
817 more efficient than the old way of doing things because it avoids levels of
818 recursion in the regex matching; (c) add a paragraph to the documentation
819 for the FullMatch() function.
821 6. The escape sequence \n was being treated as whatever was defined as
822 "newline". Not only was this contrary to the documentation, which states
823 that \n is character 10 (hex 0A), but it also went horribly wrong when
824 "newline" was defined as CRLF. This has been fixed.
826 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
827 was being set to -1 for the "end of line" case (supposedly a value that no
828 character can have). Though this value is never used (the check for end of
829 line is "zero bytes in current character"), it caused compiler complaints.
830 I've changed it to 0xffffffff.
832 8. In pcre_version.c, the version string was being built by a sequence of
833 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
834 string (as it is for production releases) called a macro with an empty
835 argument. The C standard says the result of this is undefined. The gcc
836 compiler treats it as an empty string (which was what was wanted) but it is
837 reported that Visual C gives an error. The source has been hacked around to
838 avoid this problem.
840 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
841 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
842 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
843 of them did).
845 10. Originally, pcretest opened its input and output without "b"; then I was
846 told that "b" was needed in some environments, so it was added for release
847 5.0 to both the input and output. (It makes no difference on Unix-like
848 systems.) Later I was told that it is wrong for the input on Windows. I've
849 now abstracted the modes into two macros, to make it easier to fiddle with
850 them, and removed "b" from the input mode under Windows.
852 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
854 12. Added -help and --help to pcretest as an official way of being reminded
855 of the options.
857 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
858 and pcrecpp.cc because they annoy compilers at high warning levels.
860 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
862 15. Fixed an occurrence of == in configure.ac that should have been = (shell
863 scripts are not C programs :-) and which was not noticed because it works
864 on Linux.
866 16. pcretest is supposed to handle any length of pattern and data line (as one
867 line or as a continued sequence of lines) by extending its input buffer if
868 necessary. This feature was broken for very long pattern lines, leading to
869 a string of junk being passed to pcre_compile() if the pattern was longer
870 than about 50K.
872 17. I have done a major re-factoring of the way pcre_compile() computes the
873 amount of memory needed for a compiled pattern. Previously, there was code
874 that made a preliminary scan of the pattern in order to do this. That was
875 OK when PCRE was new, but as the facilities have expanded, it has become
876 harder and harder to keep it in step with the real compile phase, and there
877 have been a number of bugs (see for example, 4 above). I have now found a
878 cunning way of running the real compile function in a "fake" mode that
879 enables it to compute how much memory it would need, while actually only
880 ever using a few hundred bytes of working memory and without too many
881 tests of the mode. This should make future maintenance and development
882 easier. A side effect of this work is that the limit of 200 on the nesting
883 depth of parentheses has been removed (though this was never a serious
884 limitation, I suspect). However, there is a downside: pcre_compile() now
885 runs more slowly than before (30% or more, depending on the pattern). I
886 hope this isn't a big issue. There is no effect on runtime performance.
888 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
889 newline (only possible for the last line of a file) and it was a
890 pattern that set a locale (followed by /Lsomething), pcretest crashed.
892 19. Added additional timing features to pcretest. (1) The -tm option now times
893 matching only, not compiling. (2) Both -t and -tm can be followed, as a
894 separate command line item, by a number that specifies the number of
895 repeats to use when timing. The default is 50000; this gives better
896 precision, but takes uncomfortably long for very large patterns.
898 20. Extended pcre_study() to be more clever in cases where a branch of a
899 subpattern has no definite first character. For example, (a*|b*)[cd] would
900 previously give no result from pcre_study(). Now it recognizes that the
901 first character must be a, b, c, or d.
903 21. There was an incorrect error "recursive call could loop indefinitely" if
904 a subpattern (or the entire pattern) that was being tested for matching an
905 empty string contained only one non-empty item after a nested subpattern.
906 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
907 incorrectly, because the \d was being skipped in the check.
909 22. The pcretest program now has a new pattern option /B and a command line
910 option -b, which is equivalent to adding /B to every pattern. This causes
911 it to show the compiled bytecode, without the additional information that
912 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
913 is the same as /B/I).
915 23. A new optimization is now able automatically to treat some sequences such
916 as a*b as a*+b. More specifically, if something simple (such as a character
917 or a simple class like \d) has an unlimited quantifier, and is followed by
918 something that cannot possibly match the quantified thing, the quantifier
919 is automatically "possessified".
921 24. A recursive reference to a subpattern whose number was greater than 39
922 went wrong under certain circumstances in UTF-8 mode. This bug could also
923 have affected the operation of pcre_study().
925 25. Realized that a little bit of performance could be had by replacing
926 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
928 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
930 27. Possessive quantifiers such as a++ were previously implemented by turning
931 them into atomic groups such as ($>a+). Now they have their own opcodes,
932 which improves performance. This includes the automatically created ones
933 from 23 above.
935 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
936 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
937 the first matched character to be a colon. This applied both to named and
938 numbered groups.
940 29. The ucpinternal.h header file was missing its idempotency #ifdef.
942 30. I was sent a "project" file called libpcre.a.dev which I understand makes
943 building PCRE on Windows easier, so I have included it in the distribution.
945 31. There is now a check in pcretest against a ridiculously large number being
946 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
947 loop, the loop is abandoned.
949 32. Forward references to subpatterns in conditions such as (?(2)...) where
950 subpattern 2 is defined later cause pcre_compile() to search forwards in
951 the pattern for the relevant set of parentheses. This search went wrong
952 when there were unescaped parentheses in a character class, parentheses
953 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
955 33. "Subroutine" calls and backreferences were previously restricted to
956 referencing subpatterns earlier in the regex. This restriction has now
957 been removed.
959 34. Added a number of extra features that are going to be in Perl 5.10. On the
960 whole, these are just syntactic alternatives for features that PCRE had
961 previously implemented using the Python syntax or my own invention. The
962 other formats are all retained for compatibility.
964 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
965 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
966 also .NET compatible.
968 (b) A recursion or subroutine call to a named group can now be defined as
969 (?&name) as well as (?P>name).
971 (c) A backreference to a named group can now be defined as \k<name> or
972 \k'name' as well as (?P=name). The new forms, as well as being in Perl
973 5.10, are also .NET compatible.
975 (d) A conditional reference to a named group can now use the syntax
976 (?(<name>) or (?('name') as well as (?(name).
978 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
979 groups (named and numbered) that are never evaluated inline, but can be
980 called as "subroutines" from elsewhere. In effect, the DEFINE condition
981 is always false. There may be only one alternative in such a group.
983 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
984 as the simple (?(R). The condition is true only if the most recent
985 recursion is that of the given number or name. It does not search out
986 through the entire recursion stack.
988 (g) The escape \gN or \g{N} has been added, where N is a positive or
989 negative number, specifying an absolute or relative reference.
991 35. Tidied to get rid of some further signed/unsigned compiler warnings and
992 some "unreachable code" warnings.
994 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
995 things, this adds five new scripts.
997 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
998 There were also incompatibilities regarding the handling of \Q..\E inside
999 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1000 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1002 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1003 matches an empty string, and forcibly breaks the loop. There were bugs in
1004 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1005 against aaaa the result was just "a" rather than "aaaa", for example. Two
1006 separate and independent bugs (that affected different cases) have been
1007 fixed.
1009 39. Refactored the code to abolish the use of different opcodes for small
1010 capturing bracket numbers. This is a tidy that I avoided doing when I
1011 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1012 The new approach is not only tidier, it makes it possible to reduce the
1013 memory needed to fix the previous bug (38).
1015 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1016 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1017 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1018 mode.
1020 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1021 report.
1023 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1024 copy construction and assignment in the C++ wrapper.
1026 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1027 couple of bugs that could have given wrong results in the "--newline=crlf"
1028 case.
1030 44. Added a number of casts and did some reorganization of signed/unsigned int
1031 variables following suggestions from Dair Grant. Also renamed the variable
1032 "this" as "item" because it is a C++ keyword.
1034 45. Arranged for dftables to add
1036 #include "pcre_internal.h"
1038 to pcre_chartables.c because without it, gcc 4.x may remove the array
1039 definition from the final binary if PCRE is built into a static library and
1040 dead code stripping is activated.
1042 46. For an unanchored pattern, if a match attempt fails at the start of a
1043 newline sequence, and the newline setting is CRLF or ANY, and the next two
1044 characters are CRLF, advance by two characters instead of one.
1047 Version 6.7 04-Jul-06
1048 ---------------------
1050 1. In order to handle tests when input lines are enormously long, pcretest has
1051 been re-factored so that it automatically extends its buffers when
1052 necessary. The code is crude, but this _is_ just a test program. The
1053 default size has been increased from 32K to 50K.
1055 2. The code in pcre_study() was using the value of the re argument before
1056 testing it for NULL. (Of course, in any sensible call of the function, it
1057 won't be NULL.)
1059 3. The memmove() emulation function in pcre_internal.h, which is used on
1060 systems that lack both memmove() and bcopy() - that is, hardly ever -
1061 was missing a "static" storage class specifier.
1063 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1064 containing an extended class (one that cannot be represented by a bitmap
1065 because it contains high-valued characters or Unicode property items, e.g.
1066 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1067 pattern, but PCRE should not loop if you do not (it no longer does).
1068 [Detail: two cases were found: (a) a repeated subpattern containing an
1069 extended class; (b) a recursive reference to a subpattern that followed a
1070 previous extended class. It wasn't skipping over the extended class
1071 correctly when UTF-8 mode was not set.]
1073 5. A negated single-character class was not being recognized as fixed-length
1074 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1075 compile error "lookbehind assertion is not fixed length".
1077 6. The RunPerlTest auxiliary script was showing an unexpected difference
1078 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1079 write a Perl script that can interpret lines of an input file either as
1080 byte characters or as UTF-8, which is what "perltest" was being required to
1081 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1082 can't do is switch easily at run time between having the "use utf8;" pragma
1083 or not. In the end, I fudged it by using the RunPerlTest script to insert
1084 "use utf8;" explicitly for the UTF-8 tests.
1086 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1087 the end of the subject string, contrary to the documentation and to what
1088 Perl does. This was true of both matching functions. Now it matches only at
1089 the start of the subject and immediately after *internal* newlines.
1091 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1092 a pointer to an int instead of a pointer to an unsigned long int. This
1093 caused problems on 64-bit systems.
1095 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1096 instance of the 'standard' template library not being so standard".
1098 10. There was no check on the number of named subpatterns nor the maximum
1099 length of a subpattern name. The product of these values is used to compute
1100 the size of the memory block for a compiled pattern. By supplying a very
1101 long subpattern name and a large number of named subpatterns, the size
1102 computation could be caused to overflow. This is now prevented by limiting
1103 the length of names to 32 characters, and the number of named subpatterns
1104 to 10,000.
1106 11. Subpatterns that are repeated with specific counts have to be replicated in
1107 the compiled pattern. The size of memory for this was computed from the
1108 length of the subpattern and the repeat count. The latter is limited to
1109 65535, but there was no limit on the former, meaning that integer overflow
1110 could in principle occur. The compiled length of a repeated subpattern is
1111 now limited to 30,000 bytes in order to prevent this.
1113 12. Added the optional facility to have named substrings with the same name.
1115 13. Added the ability to use a named substring as a condition, using the
1116 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1117 are numbers (not recommended). Forward references are permitted.
1119 14. Added forward references in named backreferences (if you see what I mean).
1121 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1122 pattern could run off the end of the subject. For example, the pattern
1123 "(?s)(.{1,5})"8 did this with the subject "ab".
1125 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1126 PCRE_CASELESS was set when matching characters that were quantified with ?
1127 or *.
1129 17. A character class other than a single negated character that had a minimum
1130 but no maximum quantifier - for example [ab]{6,} - was not handled
1131 correctly by pce_dfa_exec(). It would match only one character.
1133 18. A valid (though odd) pattern that looked like a POSIX character
1134 class but used an invalid character after [ (for example [[,abc,]]) caused
1135 pcre_compile() to give the error "Failed: internal error: code overflow" or
1136 in some cases to crash with a glibc free() error. This could even happen if
1137 the pattern terminated after [[ but there just happened to be a sequence of
1138 letters, a binary zero, and a closing ] in the memory that followed.
1140 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1141 over the years. Originally (before any Unicode support), just the bottom 8
1142 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1143 output from "man perlunicode" includes this:
1145 The regular expression compiler produces polymorphic opcodes. That
1146 is, the pattern adapts to the data and automatically switches to
1147 the Unicode character scheme when presented with Unicode data--or
1148 instead uses a traditional byte scheme when presented with byte
1149 data.
1151 Sadly, a wide octal escape does not cause a switch, and in a string with
1152 no other multibyte characters, these octal escapes are treated as before.
1153 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1154 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1155 Unicode string.
1157 I have not perpetrated such confusion in PCRE. Up till now, it took just
1158 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1159 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1160 translate to the appropriate multibyte character.
1162 29. Applied some refactoring to reduce the number of warnings from Microsoft
1163 and Borland compilers. This has included removing the fudge introduced
1164 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1165 a warning about an unused variable.
1167 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1168 characters since release 4.0, because Perl (from release 5.004) does not.
1169 [Or at least, is documented not to: some releases seem to be in conflict
1170 with the documentation.] However, when a pattern was studied with
1171 pcre_study() and all its branches started with \s, PCRE still included VT
1172 as a possible starting character. Of course, this did no harm; it just
1173 caused an unnecessary match attempt.
1175 22. Removed a now-redundant internal flag bit that recorded the fact that case
1176 dependency changed within the pattern. This was once needed for "required
1177 byte" processing, but is no longer used. This recovers a now-scarce options
1178 bit. Also moved the least significant internal flag bit to the most-
1179 significant bit of the word, which was not previously used (hangover from
1180 the days when it was an int rather than a uint) to free up another bit for
1181 the future.
1183 23. Added support for CRLF line endings as well as CR and LF. As well as the
1184 default being selectable at build time, it can now be changed at runtime
1185 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1186 specify that it is scanning data with non-default line endings.
1188 24. Changed the definition of CXXLINK to make it agree with the definition of
1189 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1191 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1192 recursions. This makes a big different to stack usage for some patterns.
1194 26. If a subpattern containing a named recursion or subroutine reference such
1195 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1196 the space required for the compiled pattern went wrong and gave too small a
1197 value. Depending on the environment, this could lead to "Failed: internal
1198 error: code overflow at offset 49" or "glibc detected double free or
1199 corruption" errors.
1201 27. Applied patches from Google (a) to support the new newline modes and (b) to
1202 advance over multibyte UTF-8 characters in GlobalReplace.
1204 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1205 difference for some implementation of PCRE in some Windows version.
1207 29. Added some extra testing facilities to pcretest:
1209 \q<number> in a data line sets the "match limit" value
1210 \Q<number> in a data line sets the "match recursion limt" value
1211 -S <number> sets the stack size, where <number> is in megabytes
1213 The -S option isn't available for Windows.
1216 Version 6.6 06-Feb-06
1217 ---------------------
1219 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1220 in pcreposix.h. I have copied the definition from pcre.h.
1222 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1223 because pcre.h is no longer a built file.
1225 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1226 not normally included in the compiled code.
1229 Version 6.5 01-Feb-06
1230 ---------------------
1232 1. When using the partial match feature with pcre_dfa_exec(), it was not
1233 anchoring the second and subsequent partial matches at the new starting
1234 point. This could lead to incorrect results. For example, with the pattern
1235 /1234/, partially matching against "123" and then "a4" gave a match.
1237 2. Changes to pcregrep:
1239 (a) All non-match returns from pcre_exec() were being treated as failures
1240 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1241 error message is output. Some extra information is given for the
1243 probably the only errors that are likely to be caused by users (by
1244 specifying a regex that has nested indefinite repeats, for instance).
1245 If there are more than 20 of these errors, pcregrep is abandoned.
1247 (b) A binary zero was treated as data while matching, but terminated the
1248 output line if it was written out. This has been fixed: binary zeroes
1249 are now no different to any other data bytes.
1251 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1252 used to set a locale for matching. The --locale=xxxx long option has
1253 been added (no short equivalent) to specify a locale explicitly on the
1254 pcregrep command, overriding the environment variables.
1256 (d) When -B was used with -n, some line numbers in the output were one less
1257 than they should have been.
1259 (e) Added the -o (--only-matching) option.
1261 (f) If -A or -C was used with -c (count only), some lines of context were
1262 accidentally printed for the final match.
1264 (g) Added the -H (--with-filename) option.
1266 (h) The combination of options -rh failed to suppress file names for files
1267 that were found from directory arguments.
1269 (i) Added the -D (--devices) and -d (--directories) options.
1271 (j) Added the -F (--fixed-strings) option.
1273 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1275 (l) Added the --colo(u)r option.
1277 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1278 is not present by default.
1280 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1281 items such as (?R) or (?1), when the recursion could match a number of
1282 alternatives. If it matched one of the alternatives, but subsequently,
1283 outside the recursion, there was a failure, the code tried to back up into
1284 the recursion. However, because of the way PCRE is implemented, this is not
1285 possible, and the result was an incorrect result from the match.
1287 In order to prevent this happening, the specification of recursion has
1288 been changed so that all such subpatterns are automatically treated as
1289 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1291 4. I had overlooked the fact that, in some locales, there are characters for
1292 which isalpha() is true but neither isupper() nor islower() are true. In
1293 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1294 and ordfeminine) are like this. This affected the treatment of \w and \W
1295 when they appeared in character classes, but not when they appeared outside
1296 a character class. The bit map for "word" characters is now created
1297 separately from the results of isalnum() instead of just taking it from the
1298 upper, lower, and digit maps. (Plus the underscore character, of course.)
1300 5. The above bug also affected the handling of POSIX character classes such as
1301 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1302 permanent tables. Instead, the bit maps for such a class were previously
1303 created as the appropriate unions of the upper, lower, and digit bitmaps.
1304 Now they are created by subtraction from the [[:word:]] class, which has
1305 its own bitmap.
1307 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1308 It is created by subtracting the vertical space characters (\x09, \x0a,
1309 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1310 subtraction was done in the overall bitmap for a character class, meaning
1311 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1312 be recognized. This bug has been fixed.
1314 7. Patches from the folks at Google:
1316 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1317 real life, but is still worth protecting against".
1319 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1320 regular expressions".
1322 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1323 have it.
1325 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1326 "configure" and the latter not, in order to fix a problem somebody had
1327 with compiling the Arg class on HP-UX.
1329 (e) Improve the error-handling of the C++ wrapper a little bit.
1331 (f) New tests for checking recursion limiting.
1333 8. The pcre_memmove() function, which is used only if the environment does not
1334 have a standard memmove() function (and is therefore rarely compiled),
1335 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1336 returning a result (though PCRE never actually uses the result).
1338 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1339 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1340 returned instead of calling malloc() with an overflowing number that would
1341 most likely cause subsequent chaos.
1343 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1345 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1346 with this option is matched, the nmatch and pmatch options of regexec() are
1347 ignored.
1349 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1350 provided in case anyone wants to the the POSIX interface with UTF-8
1351 strings.
1353 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1354 C++ linking (needed for some HP-UX environments).
1356 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1357 (unused parameter) and in the pcre_printint() function (omitted "default"
1358 switch label when the default is to do nothing).
1360 15. Added some code to make it possible, when PCRE is compiled as a C++
1361 library, to replace subject pointers for pcre_exec() with a smart pointer
1362 class, thus making it possible to process discontinuous strings.
1364 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1365 much the same function. They were added by different people who were trying
1366 to make PCRE easy to compile on non-Unix systems. It has been suggested
1367 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1368 for compiling on Windows systems. I have therefore replaced it with
1369 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1370 defaults to "extern" for C or "extern C" for C++, which works fine on
1371 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1372 SCOPE with something explicit in config.h. In addition:
1374 (a) pcreposix.h still had just "extern" instead of either of these macros;
1375 I have replaced it with PCRE_DATA_SCOPE.
1377 (b) Functions such as _pcre_xclass(), which are internal to the library,
1378 but external in the C sense, all had PCRE_EXPORT in their definitions.
1379 This is apparently wrong for the Windows case, so I have removed it.
1380 (It makes no difference on Unix-like systems.)
1382 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1383 of recursive calls to match(). This is different to MATCH_LIMIT because
1384 that limits the total number of calls to match(), not all of which increase
1385 the depth of recursion. Limiting the recursion depth limits the amount of
1386 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1387 when PCRE is compiled, and changed at run time. A patch from Google adds
1388 this functionality to the C++ interface.
1390 18. Changes to the handling of Unicode character properties:
1392 (a) Updated the table to Unicode 4.1.0.
1394 (b) Recognize characters that are not in the table as "Cn" (undefined).
1396 (c) I revised the way the table is implemented to a much improved format
1397 which includes recognition of ranges. It now supports the ranges that
1398 are defined in UnicodeData.txt, and it also amalgamates other
1399 characters into ranges. This has reduced the number of entries in the
1400 table from around 16,000 to around 3,000, thus reducing its size
1401 considerably. I realized I did not need to use a tree structure after
1402 all - a binary chop search is just as efficient. Having reduced the
1403 number of entries, I extended their size from 6 bytes to 8 bytes to
1404 allow for more data.
1406 (d) Added support for Unicode script names via properties such as \p{Han}.
1408 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1409 matching that character.
1411 20. When matching a repeated Unicode property with a minimum greater than zero,
1412 (for example \pL{2,}), PCRE could look past the end of the subject if it
1413 reached it while seeking the minimum number of characters. This could
1414 happen only if some of the characters were more than one byte long, because
1415 there is a check for at least the minimum number of bytes.
1417 21. Refactored the implementation of \p and \P so as to be more general, to
1418 allow for more different types of property in future. This has changed the
1419 compiled form incompatibly. Anybody with saved compiled patterns that use
1420 \p or \P will have to recompile them.
1422 22. Added "Any" and "L&" to the supported property types.
1424 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1425 but give a compile time error if the value is greater than 0xff.
1427 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1428 accidentally not being installed or uninstalled.
1430 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1431 made were to insert the current release number. This seemed silly, because
1432 it made things harder for people building PCRE on systems that don't run
1433 "configure". I have turned pcre.h into a distributed file, no longer built
1434 by "configure", with the version identification directly included. There is
1435 no longer a pcre.h.in file.
1437 However, this change necessitated a change to the pcre-config script as
1438 well. It is built from pcre-config.in, and one of the substitutions was the
1439 release number. I have updated configure.ac so that ./configure now finds
1440 the release number by grepping pcre.h.
1442 26. Added the ability to run the tests under valgrind.
1445 Version 6.4 05-Sep-05
1446 ---------------------
1448 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1449 "--" to be printed when multiple files were scanned, even when none of the
1450 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1451 consider it to be a bug, and have restored the previous behaviour.
1453 2. A couple of code tidies to get rid of compiler warnings.
1455 3. The pcretest program used to cheat by referring to symbols in the library
1456 whose names begin with _pcre_. These are internal symbols that are not
1457 really supposed to be visible externally, and in some environments it is
1458 possible to suppress them. The cheating is now confined to including
1459 certain files from the library's source, which is a bit cleaner.
1461 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1462 file's purpose clearer.
1464 5. Reorganized pcre_ucp_findchar().
1467 Version 6.3 15-Aug-05
1468 ---------------------
1470 1. The file libpcre.pc.in did not have general read permission in the tarball.
1472 2. There were some problems when building without C++ support:
1474 (a) If C++ support was not built, "make install" and "make test" still
1475 tried to test it.
1477 (b) There were problems when the value of CXX was explicitly set. Some
1478 changes have been made to try to fix these, and ...
1480 (c) --disable-cpp can now be used to explicitly disable C++ support.
1482 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1483 backslash in a target when C++ was disabled. This confuses some
1484 versions of "make", apparently. Using an intermediate variable solves
1485 this. (Same for CPP_LOBJ.)
1487 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1488 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1489 necessary on certain architectures.
1491 4. Added a setting of -export-symbols-regex to the link command to remove
1492 those symbols that are exported in the C sense, but actually are local
1493 within the library, and not documented. Their names all begin with
1494 "_pcre_". This is not a perfect job, because (a) we have to except some
1495 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1496 available (and never for static libraries). I have made a note to try to
1497 find a way round (a) in the future.
1500 Version 6.2 01-Aug-05
1501 ---------------------
1503 1. There was no test for integer overflow of quantifier values. A construction
1504 such as {1111111111111111} would give undefined results. What is worse, if
1505 a minimum quantifier for a parenthesized subpattern overflowed and became
1506 negative, the calculation of the memory size went wrong. This could have
1507 led to memory overwriting.
1509 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1511 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1512 operating environments where this matters.
1514 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1515 PCRE options from within the C++ wrapper.
1517 5. Named capturing subpatterns were not being correctly counted when a pattern
1518 was compiled. This caused two problems: (a) If there were more than 100
1519 such subpatterns, the calculation of the memory needed for the whole
1520 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1521 back references of the form \12, where the number was greater than 9, were
1522 not recognized as back references, even though there were sufficient
1523 previous subpatterns.
1525 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1526 versions of gcc, e.g. 2.95.4.
1529 Version 6.1 21-Jun-05
1530 ---------------------
1532 1. There was one reference to the variable "posix" in pcretest.c that was not
1533 surrounded by "#if !defined NOPOSIX".
1535 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1536 the cross-check on the old pcre_info() function, for the benefit of the
1537 cut-down version of PCRE that is currently imported into Exim.
1539 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1540 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1541 patterns that start (?i) but I don't think that matters. The use of (?i) is
1542 just an example; this all applies to the other options as well.
1544 4. Since libtool seems to echo the compile commands it is issuing, the output
1545 from "make" can be reduced a bit by putting "@" in front of each libtool
1546 compile command.
1548 5. Patch from the folks at Google for configure.in to be a bit more thorough
1549 in checking for a suitable C++ installation before trying to compile the
1550 C++ stuff. This should fix a reported problem when a compiler was present,
1551 but no suitable headers.
1553 6. The man pages all had just "PCRE" as their title. I have changed them to
1554 be the relevant file name. I have also arranged that these names are
1555 retained in the file doc/pcre.txt, which is a concatenation in text format
1556 of all the man pages except the little individual ones for each function.
1558 7. The NON-UNIX-USE file had not been updated for the different set of source
1559 files that come with release 6. I also added a few comments about the C++
1560 wrapper.
1563 Version 6.0 07-Jun-05
1564 ---------------------
1566 1. Some minor internal re-organization to help with my DFA experiments.
1568 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1569 didn't matter for the library itself when fully configured, but did matter
1570 when compiling without UCP support, or within Exim, where the ucp files are
1571 not imported.
1573 3. Refactoring of the library code to split up the various functions into
1574 different source modules. The addition of the new DFA matching code (see
1575 below) to a single monolithic source would have made it really too
1576 unwieldy, quite apart from causing all the code to be include in a
1577 statically linked application, when only some functions are used. This is
1578 relevant even without the DFA addition now that patterns can be compiled in
1579 one application and matched in another.
1581 The downside of splitting up is that there have to be some external
1582 functions and data tables that are used internally in different modules of
1583 the library but which are not part of the API. These have all had their
1584 names changed to start with "_pcre_" so that they are unlikely to clash
1585 with other external names.
1587 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1588 a different (DFA) algorithm. Although it is slower than the original
1589 function, it does have some advantages for certain types of matching
1590 problem.
1592 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1593 including restarting after a partial match.
1595 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1596 defined when compiling for Windows was sent to me. I have put it into the
1597 code, though I have no means of testing or verifying it.
1599 7. Added the pcre_refcount() auxiliary function.
1601 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1602 match before or at the first newline in the subject string. In pcretest,
1603 the /f option on a pattern can be used to set this.
1605 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1606 would behave wrongly. This has been present in PCRE since release 4.0.
1608 10. A number of changes to the pcregrep command:
1610 (a) Refactored how -x works; insert ^(...)$ instead of setting
1611 PCRE_ANCHORED and checking the length, in preparation for adding
1612 something similar for -w.
1614 (b) Added the -w (match as a word) option.
1616 (c) Refactored the way lines are read and buffered so as to have more
1617 than one at a time available.
1619 (d) Implemented a pcregrep test script.
1621 (e) Added the -M (multiline match) option. This allows patterns to match
1622 over several lines of the subject. The buffering ensures that at least
1623 8K, or the rest of the document (whichever is the shorter) is available
1624 for matching (and similarly the previous 8K for lookbehind assertions).
1626 (f) Changed the --help output so that it now says
1628 -w, --word-regex(p)
1630 instead of two lines, one with "regex" and the other with "regexp"
1631 because that confused at least one person since the short forms are the
1632 same. (This required a bit of code, as the output is generated
1633 automatically from a table. It wasn't just a text change.)
1635 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1636 option but starts with a hyphen. Could be a pattern or a path name
1637 starting with a hyphen, for instance.
1639 (h) "-" can be given as a file name to represent stdin.
1641 (i) When file names are being printed, "(standard input)" is used for
1642 the standard input, for compatibility with GNU grep. Previously
1643 "<stdin>" was used.
1645 (j) The option --label=xxx can be used to supply a name to be used for
1646 stdin when file names are being printed. There is no short form.
1648 (k) Re-factored the options decoding logic because we are going to add
1649 two more options that take data. Such options can now be given in four
1650 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1652 (l) Added the -A, -B, and -C options for requesting that lines of context
1653 around matches be printed.
1655 (m) Added the -L option to print the names of files that do not contain
1656 any matching lines, that is, the complement of -l.
1658 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1659 continue to scan other files.
1661 (o) The -s option was incorrectly implemented. For compatibility with other
1662 greps, it now suppresses the error message for a non-existent or non-
1663 accessible file (but not the return code). There is a new option called
1664 -q that suppresses the output of matching lines, which was what -s was
1665 previously doing.
1667 (p) Added --include and --exclude options to specify files for inclusion
1668 and exclusion when recursing.
1670 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1671 Hopefully, it now does.
1673 12. Missing cast in pcre_study().
1675 13. Added an "uninstall" target to the makefile.
1677 14. Replaced "extern" in the function prototypes in Makefile.in with
1678 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1679 world, but is set differently for Windows.
1681 15. Added a second compiling function called pcre_compile2(). The only
1682 difference is that it has an extra argument, which is a pointer to an
1683 integer error code. When there is a compile-time failure, this is set
1684 non-zero, in addition to the error test pointer being set to point to an
1685 error message. The new argument may be NULL if no error number is required
1686 (but then you may as well call pcre_compile(), which is now just a
1687 wrapper). This facility is provided because some applications need a
1688 numeric error indication, but it has also enabled me to tidy up the way
1689 compile-time errors are handled in the POSIX wrapper.
1691 16. Added VPATH=.libs to the makefile; this should help when building with one
1692 prefix path and installing with another. (Or so I'm told by someone who
1693 knows more about this stuff than I do.)
1695 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1696 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1697 match everything, including newlines. This is not POSIX-compatible, but
1698 somebody wanted the feature. From pcretest it can be activated by using
1699 both the P and the s flags.
1701 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1703 19. libpcre.pc was being incorrectly installed as executable.
1705 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1706 it now also looks for '\r' so that it will work unmodified on Windows.
1708 21. Added Google's contributed C++ wrapper to the distribution.
1710 22. Added some untidy missing memory free() calls in pcretest, to keep
1711 Electric Fence happy when testing.
1715 Version 5.0 13-Sep-04
1716 ---------------------
1718 1. Internal change: literal characters are no longer packed up into items
1719 containing multiple characters in a single byte-string. Each character
1720 is now matched using a separate opcode. However, there may be more than one
1721 byte in the character in UTF-8 mode.
1723 2. The pcre_callout_block structure has two new fields: pattern_position and
1724 next_item_length. These contain the offset in the pattern to the next match
1725 item, and its length, respectively.
1727 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1728 insertion of callouts before each pattern item. Added the /C option to
1729 pcretest to make use of this.
1731 4. On the advice of a Windows user, the lines
1733 #if defined(_WIN32) || defined(WIN32)
1734 _setmode( _fileno( stdout ), 0x8000 );
1735 #endif /* defined(_WIN32) || defined(WIN32) */
1737 have been added to the source of pcretest. This apparently does useful
1738 magic in relation to line terminators.
1740 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1741 for the benefit of those environments where the "b" makes a difference.
1743 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1744 to know about it. I have put a hack into configure.in that adds in code
1745 to set GCC=yes if CC=icc. This seems to end up at a point in the
1746 generated configure script that is early enough to affect the setting of
1747 compiler options, which is what is needed, but I have no means of testing
1748 whether it really works. (The user who reported this had patched the
1749 generated configure script, which of course I cannot do.)
1751 LATER: After change 22 below (new libtool files), the configure script
1752 seems to know about icc (and also ecc). Therefore, I have commented out
1753 this hack in configure.in.
1755 7. Added support for pkg-config (2 patches were sent in).
1757 8. Negated POSIX character classes that used a combination of internal tables
1758 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1759 [[:^ascii]]. Typically, they would match almost any characters. The other
1760 POSIX classes were not broken in this way.
1762 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1763 to find the match, as PCRE was deluded into thinking that the match had to
1764 start at the start point or following a newline. The same bug applied to
1765 patterns with negative forward assertions or any backward assertions
1766 preceding ".*" at the start, unless the pattern required a fixed first
1767 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1769 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1770 starting at the last subject character, bytes beyond the end of the subject
1771 string were read.
1773 11. Renamed the variable "class" as "classbits" to make life easier for C++
1774 users. (Previously there was a macro definition, but it apparently wasn't
1775 enough.)
1777 12. Added the new field "tables" to the extra data so that tables can be passed
1778 in at exec time, or the internal tables can be re-selected. This allows
1779 a compiled regex to be saved and re-used at a later time by a different
1780 program that might have everything at different addresses.
1782 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1783 -R library as well as a -L library.
1785 14. The debugging options of pcretest (-d on the command line or D on a
1786 pattern) showed incorrect output for anything following an extended class
1787 that contained multibyte characters and which was followed by a quantifier.
1789 15. Added optional support for general category Unicode character properties
1790 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1791 support. It adds about 90K to the size of the library. The meanings of the
1792 inbuilt class escapes such as \d and \s have NOT been changed.
1794 16. Updated pcredemo.c to include calls to free() to release the memory for the
1795 compiled pattern.
1797 17. The generated file chartables.c was being created in the source directory
1798 instead of in the building directory. This caused the build to fail if the
1799 source directory was different from the building directory, and was
1800 read-only.
1802 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1803 file. No doubt somebody will tell me if they don't make sense... Also added
1804 Dan Mooney's comments about building on OpenVMS.
1806 19. Added support for partial matching via the PCRE_PARTIAL option for
1807 pcre_exec() and the \P data escape in pcretest.
1809 20. Extended pcretest with 3 new pattern features:
1811 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1812 write the compiled pattern to the file whose name is "rest-of-line".
1813 This is a straight binary dump of the data, with the saved pointer to
1814 the character tables forced to be NULL. The study data, if any, is
1815 written too. After writing, pcretest reads a new pattern.
1817 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1818 compiled pattern from the given file. There must not be any
1819 occurrences of "<" in the file name (pretty unlikely); if there are,
1820 pcretest will instead treat the initial "<" as a pattern delimiter.
1821 After reading in the pattern, pcretest goes on to read data lines as
1822 usual.
1824 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1825 and 16-bit fields in a compiled pattern, to simulate a pattern that
1826 was compiled on a host of opposite endianness.
1828 21. The pcre-exec() function can now cope with patterns that were compiled on
1829 hosts of opposite endianness, with this restriction:
1831 As for any compiled expression that is saved and used later, the tables
1832 pointer field cannot be preserved; the extra_data field in the arguments
1833 to pcre_exec() should be used to pass in a tables address if a value
1834 other than the default internal tables were used at compile time.
1836 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1837 now diagnosed as an error. Previously, most of the time, a negative number
1838 would have been treated as zero, but if in addition "ovector" was passed as
1839 NULL, a crash could occur.
1841 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1842 new versions from the libtool 1.5 distribution (the last one is a copy of
1843 a file called libtool.m4). This seems to have fixed the need to patch
1844 "configure" to support Darwin 1.3 (which I used to do). However, I still
1845 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1846 workstation).
1848 24. Changed the PCRE licence to be the more standard "BSD" licence.
1851 Version 4.5 01-Dec-03
1852 ---------------------
1854 1. There has been some re-arrangement of the code for the match() function so
1855 that it can be compiled in a version that does not call itself recursively.
1856 Instead, it keeps those local variables that need separate instances for
1857 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1858 needs to "recurse". Keeping track of where control must go is done by means
1859 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1860 hide most of the details from the main code, and operates only if
1861 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1862 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1863 operating.
1865 To make it easier for callers to provide specially tailored get/free
1866 functions for this usage, two new functions, pcre_stack_malloc, and
1867 pcre_stack_free, are used. They are always called in strict stacking order,
1868 and the size of block requested is always the same.
1870 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1871 PCRE has been compiled to use the stack or the heap for recursion. The
1872 -C option of pcretest uses this to show which version is compiled.
1874 A new data escape \S, is added to pcretest; it causes the amounts of store
1875 obtained and freed by both kinds of malloc/free at match time to be added
1876 to the output.
1878 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1879 what's available on my current Linux desktop machine.
1881 3. When matching a UTF-8 string, the test for a valid string at the start has
1882 been extended. If start_offset is not zero, PCRE now checks that it points
1883 to a byte that is the start of a UTF-8 character. If not, it returns
1884 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1885 this is necessary because there may be backward assertions in the pattern.
1886 When matching the same subject several times, it may save resources to use
1887 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1889 4. The code for checking the validity of UTF-8 strings has been tightened so
1890 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1891 containing "overlong sequences".
1893 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1894 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1895 should have been "&", but it just so happened that all the cases this let
1896 through by mistake were picked up later in the function.
1898 6. I had used a variable called "isblank" - this is a C99 function, causing
1899 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1901 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1902 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1903 so the version is shown; (c) stop "make test" echoing "./RunTest".
1905 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1907 9. The prototype for memmove() for systems that don't have it was using
1908 size_t, but the inclusion of the header that defines size_t was later. I've
1909 moved the #includes for the C headers earlier to avoid this.
1911 10. Added some adjustments to the code to make it easier to compiler on certain
1912 special systems:
1914 (a) Some "const" qualifiers were missing.
1915 (b) Added the macro EXPORT before all exported functions; by default this
1916 is defined to be empty.
1917 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1918 that it reads its output file name as an argument instead of writing
1919 to the standard output and assuming this can be redirected.
1921 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1922 class containing characters with values greater than 255, PCRE compilation
1923 went into a loop.
1925 12. A recursive reference to a subpattern that was within another subpattern
1926 that had a minimum quantifier of zero caused PCRE to crash. For example,
1927 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1928 recursion. If the recursively-called subpattern itself had a zero repeat,
1929 that was OK.
1931 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1932 buffer into which it was copied (for escape processing) was still set at
1933 1024, so long lines caused crashes.
1935 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1936 "internal error: code overflow...". This applied to any character class
1937 that was followed by a possessive quantifier.
1939 15. Modified the Makefile to add libpcre.la as a prerequisite for
1940 libpcreposix.la because I was told this is needed for a parallel build to
1941 work.
1943 16. If a pattern that contained .* following optional items at the start was
1944 studied, the wrong optimizing data was generated, leading to matching
1945 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1946 matching string must start with a or b or c. The correct conclusion for
1947 this pattern is that a match can start with any character.
1950 Version 4.4 13-Aug-03
1951 ---------------------
1953 1. In UTF-8 mode, a character class containing characters with values between
1954 127 and 255 was not handled correctly if the compiled pattern was studied.
1955 In fixing this, I have also improved the studying algorithm for such
1956 classes (slightly).
1958 2. Three internal functions had redundant arguments passed to them. Removal
1959 might give a very teeny performance improvement.
1961 3. Documentation bug: the value of the capture_top field in a callout is *one
1962 more than* the number of the hightest numbered captured substring.
1964 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1965 in incorrectly linking with a previously installed version. They now link
1966 explicitly with libpcre.la.
1968 5. configure.in no longer needs to recognize Cygwin specially.
1970 6. A problem in pcre.in for Windows platforms is fixed.
1972 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1973 pcretest, it used to include the size of the study block as part of its
1974 output. Unfortunately, the structure contains a field that has a different
1975 size on different hardware architectures. This meant that the tests that
1976 showed this size failed. As the block is currently always of a fixed size,
1977 this information isn't actually particularly useful in pcretest output, so
1978 I have just removed it.
1980 8. Three pre-processor statements accidentally did not start in column 1.
1981 Sadly, there are *still* compilers around that complain, even though
1982 standard C has not required this for well over a decade. Sigh.
1984 9. In pcretest, the code for checking callouts passed small integers in the
1985 callout_data field, which is a void * field. However, some picky compilers
1986 complained about the casts involved for this on 64-bit systems. Now
1987 pcretest passes the address of the small integer instead, which should get
1988 rid of the warnings.
1990 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1991 both compile and run time, and gives an error if an invalid UTF-8 sequence
1992 is found. There is a option for disabling this check in cases where the
1993 string is known to be correct and/or the maximum performance is wanted.
1995 11. In response to a bug report, I changed one line in Makefile.in from
1997 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1998 to
1999 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2001 to look similar to other lines, but I have no way of telling whether this
2002 is the right thing to do, as I do not use Windows. No doubt I'll get told
2003 if it's wrong...
2006 Version 4.3 21-May-03
2007 ---------------------
2009 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2010 Makefile.
2012 2. Some refactoring to improve the quality of the code:
2014 (i) The utf8_table... variables are now declared "const".
2016 (ii) The code for \cx, which used the "case flipping" table to upper case
2017 lower case letters, now just substracts 32. This is ASCII-specific,
2018 but the whole concept of \cx is ASCII-specific, so it seems
2019 reasonable.
2021 (iii) PCRE was using its character types table to recognize decimal and
2022 hexadecimal digits in the pattern. This is silly, because it handles
2023 only 0-9, a-f, and A-F, but the character types table is locale-
2024 specific, which means strange things might happen. A private
2025 table is now used for this - though it costs 256 bytes, a table is
2026 much faster than multiple explicit tests. Of course, the standard
2027 character types table is still used for matching digits in subject
2028 strings against \d.
2030 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2031 ending in _t are). So I've renamed it as ESC_tee.
2033 3. The first argument for regexec() in the POSIX wrapper should have been
2034 defined as "const".
2036 4. Changed pcretest to use malloc() for its buffers so that they can be
2037 Electric Fenced for debugging.
2039 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2040 to read one or more bytes before the start of the subject string. Often this
2041 had no effect on PCRE's behaviour, but in some circumstances it could
2042 provoke a segmentation fault.
2044 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2045 to try to read one or more bytes before the start of the subject string.
2047 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2048 UTF-8 support could misbehave in various ways if the subject string
2049 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2050 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2051 back over UTF-8 characters.)
2054 Version 4.2 14-Apr-03
2055 ---------------------
2057 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2059 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2060 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2061 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2062 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2063 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2065 Note: automatic setting of the BUILD variables is not yet working
2066 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2067 compile-time but not at link-time
2068 [LINK]: use for linking executables only
2069 make different versions for Windows and non-Windows
2070 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2071 libraries
2072 [LINK_FOR_BUILD]: new variable
2073 [OBJEXT]: use throughout
2074 [EXEEXT]: use throughout
2075 <winshared>: new target
2076 <wininstall>: new target
2077 <dftables.o>: use native compiler
2078 <dftables>: use native linker
2079 <install>: handle Windows platform correctly
2080 <clean>: ditto
2081 <check>: ditto
2082 copy DLL to top builddir before testing
2084 As part of these changes, -no-undefined was removed again. This was reported
2085 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2086 in any case.
2088 3. Some tidies to get rid of compiler warnings:
2090 . In the match_data structure, match_limit was an unsigned long int, whereas
2091 match_call_count was an int. I've made them both unsigned long ints.
2093 . In pcretest the fact that a const uschar * doesn't automatically cast to
2094 a void * provoked a warning.
2096 . Turning on some more compiler warnings threw up some "shadow" variables
2097 and a few more missing casts.
2099 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2100 option, a class that contained a single character with a value between 128
2101 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2103 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2104 option, a class that contained several characters, but with at least one
2105 whose value was between 128 and 255 caused PCRE to crash.
2108 Version 4.1 12-Mar-03
2109 ---------------------
2111 1. Compiling with gcc -pedantic found a couple of places where casts were
2112 needed, and a string in dftables.c that was longer than standard compilers are
2113 required to support.
2115 2. Compiling with Sun's compiler found a few more places where the code could
2116 be tidied up in order to avoid warnings.
2118 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2119 first of these names is deprecated in the latest Autoconf in favour of the name
2120 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2121 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2122 analogy I have changed it to CFLAGS_FOR_BUILD.
2124 4. Added -no-undefined to the linking command in the Makefile, because this is
2125 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2126 linking step for the pcreposix library.
2128 5. PCRE was failing to diagnose the case of two named groups with the same
2129 name.
2131 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2132 literal character that is needed in the subject for a match, and scans along to
2133 ensure that it is present before embarking on the full matching process. This
2134 saves time in cases of nested unlimited repeats that are never going to match.
2135 Problem: the scan can take a lot of time if the subject is very long (e.g.
2136 megabytes), thus penalizing straightforward matches. It is now done only if the
2137 amount of subject to be scanned is less than 1000 bytes.
2139 7. A lesser problem with the same optimization is that it was recording the
2140 first character of an anchored pattern as "needed", thus provoking a search
2141 right along the subject, even when the first match of the pattern was going to
2142 fail. The "needed" character is now not set for anchored patterns, unless it
2143 follows something in the pattern that is of non-fixed length. Thus, it still
2144 fulfils its original purpose of finding quick non-matches in cases of nested
2145 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2148 Version 4.0 17-Feb-03
2149 ---------------------
2151 1. If a comment in an extended regex that started immediately after a meta-item
2152 extended to the end of string, PCRE compiled incorrect data. This could lead to
2153 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2155 2. Moved to autoconf 2.53 and libtool 1.4.2.
2157 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2158 the special perltest8 script is no longer needed - all the tests can be run
2159 from a single perltest script.
2161 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2162 by \s. It has now been removed in PCRE. This means it isn't recognized as
2163 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2164 class [:space:] *does* include VT, thereby creating a mess.
2166 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2167 space and tab.
2169 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2170 its new features into the main test script, reducing the number of scripts.
2172 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2173 were backward compatible, and made the (?i) apply to the whole pattern, as if
2174 /i were given. Now it behaves more logically, and applies the option setting
2175 only to what follows. PCRE has been changed to follow suit. However, if it
2176 finds options settings right at the start of the pattern, it extracts them into
2177 the global options, as before. Thus, they show up in the info data.
2179 8. Added support for the \Q...\E escape sequence. Characters in between are
2180 treated as literals. This is slightly different from Perl in that $ and @ are
2181 also handled as literals inside the quotes. In Perl, they will cause variable
2182 interpolation. Note the following examples:
2184 Pattern PCRE matches Perl matches
2186 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2187 \Qabc\$xyz\E abc\$xyz abc\$xyz
2188 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2190 For compatibility with Perl, \Q...\E sequences are recognized inside character
2191 classes as well as outside them.
2193 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2194 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2195 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2196 signed/unsigned warnings.
2198 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2199 option for pcretest, so I've replaced it by a simple function that does just
2200 that job.
2202 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2203 "pcregrep -".
2205 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2206 Java package. This provides some syntactic sugar for simple cases of what my
2207 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2208 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2209 item, you can use this simplified notation. Note that only makes sense with
2210 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2211 greediness, whatever the setting of the PCRE_UNGREEDY option.
2213 13. A change of greediness default within a pattern was not taking effect at
2214 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2215 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2216 was abstracted outside.
2218 14. PCRE now supports the \G assertion. It is true when the current matching
2219 position is at the start point of the match. This differs from \A when the
2220 starting offset is non-zero. Used with the /g option of pcretest (or similar
2221 code), it works in the same way as it does for Perl's /g option. If all
2222 alternatives of a regex begin with \G, the expression is anchored to the start
2223 match position, and the "anchored" flag is set in the compiled expression.
2225 15. Some bugs concerning the handling of certain option changes within patterns
2226 have been fixed. These applied to options other than (?ims). For example,
2227 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2228 been the other way round. Some of this was related to change 7 above.
2230 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2231 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2232 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2233 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2235 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2236 mode. Unlike ".", it always matches newline, whatever the setting of
2237 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2238 assertions. Perl allows it, but it doesn't (in general) work because it can't
2239 calculate the length of the lookbehind. At least, that's the case for Perl
2240 5.8.0 - I've been told they are going to document that it doesn't work in
2241 future.
2243 18. Added an error diagnosis for escapes that PCRE does not support: these are
2244 \L, \l, \N, \P, \p, \U, \u, and \X.
2246 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2247 reading past the end of the pattern in cases such as /[abcd/.
2249 20. PCRE was getting more memory than necessary for patterns with classes that
2250 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2252 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2253 compiling PCRE for use with Virtual Pascal.
2255 22. Small fix to the Makefile to make it work properly if the build is done
2256 outside the source tree.
2258 23. Added a new extension: a condition to go with recursion. If a conditional
2259 subpattern starts with (?(R) the "true" branch is used if recursion has
2260 happened, whereas the "false" branch is used only at the top level.
2262 24. When there was a very long string of literal characters (over 255 bytes
2263 without UTF support, over 250 bytes with UTF support), the computation of how
2264 much memory was required could be incorrect, leading to segfaults or other
2265 strange effects.
2267 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2268 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2269 there was a subsequent back reference to those brackets. This meant that, for
2270 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2271 possible to check for precisely this case. All we can do is abandon the
2272 optimization if .* occurs inside capturing brackets when there are any back
2273 references whatsoever. (See below for a better fix that came later.)
2275 26. The handling of the optimization for finding the first character of a
2276 non-anchored pattern, and for finding a character that is required later in the
2277 match were failing in some cases. This didn't break the matching; it just
2278 failed to optimize when it could. The way this is done has been re-implemented.
2280 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2282 28. Added a new feature that provides some of the functionality that Perl
2283 provides with (?{...}). The facility is termed a "callout". The way it is done
2284 in PCRE is for the caller to provide an optional function, by setting
2285 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2286 global variable. By default it is unset, which disables all calling out. To get
2287 the function called, the regex must include (?C) at appropriate points. This
2288 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2289 This provides a means of identifying different callout points. When PCRE
2290 reaches such a point in the regex, if pcre_callout has been set, the external
2291 function is called. It is provided with data in a structure called
2292 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2293 matching continues; if it returns a non-zero value, the match at the current
2294 point fails. However, backtracking will occur if possible. [This was changed
2295 later and other features added - see item 49 below.]
2297 29. pcretest is upgraded to test the callout functionality. It provides a
2298 callout function that displays information. By default, it shows the start of
2299 the match and the current position in the text. There are some new data escapes
2300 to vary what happens:
2302 \C+ in addition, show current contents of captured substrings
2303 \C- do not supply a callout function
2304 \C!n return 1 when callout number n is reached
2305 \C!n!m return 1 when callout number n is reached for the mth time
2307 30. If pcregrep was called with the -l option and just a single file name, it
2308 output "<stdin>" if a match was found, instead of the file name.
2310 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2311 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2312 pcre_exec(). This saves a malloc/free per call. The default value of
2313 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2314 when configuring.
2316 32. The default maximum size of a compiled pattern is 64K. There have been a
2317 few cases of people hitting this limit. The code now uses macros to handle the
2318 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2319 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2320 configuring. Tests 2 and 5 work only with 2-byte links because they output
2321 debugging information about compiled patterns.
2323 33. Internal code re-arrangements:
2325 (a) Moved the debugging function for printing out a compiled regex into
2326 its own source file (printint.c) and used #include to pull it into
2327 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2328 separate copies.
2330 (b) Defined the list of op-code names for debugging as a macro in
2331 internal.h so that it is next to the definition of the opcodes.
2333 (c) Defined a table of op-code lengths for simpler skipping along compiled
2334 code. This is again a macro in internal.h so that it is next to the
2335 definition of the opcodes.
2337 34. Added support for recursive calls to individual subpatterns, along the
2338 lines of Robin Houston's patch (but implemented somewhat differently).
2340 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2341 allow it to read and process whole directories in Win32. This code was
2342 contributed by Lionel Fourquaux; it has not been tested by me.
2344 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2345 used to name a group. Names consist of alphanumerics and underscores, and must
2346 be unique. Back references use the syntax (?P=name) and recursive calls use
2347 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2348 numbers. The function pcre_fullinfo() can be used after compilation to extract
2349 a name/number map. There are three relevant calls:
2351 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2352 PCRE_INFO_NAMECOUNT yields the number of entries
2353 PCRE_INFO_NAMETABLE yields a pointer to the map.
2355 The map is a vector of fixed-size entries. The size of each entry depends on
2356 the length of the longest name used. The first two bytes of each entry are the
2357 group number, most significant byte first. There follows the corresponding
2358 name, zero terminated. The names are in alphabetical order.
2360 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2361 case instead of 255. Making it the same both with and without UTF-8 support
2362 means that the same test output works with both.
2364 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2365 calling malloc() with a zero argument.
2367 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2368 optimization. I've improved things by keeping a bitmap of backreferences with
2369 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2370 fact referenced, the optimization can be applied. It is unlikely that a
2371 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2372 the match to follow \n) will appear inside brackets with a number greater than
2373 31, but if it does, any back reference > 31 suppresses the optimization.
2375 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2376 of disabling numbered capturing parentheses. Any opening parenthesis that is
2377 not followed by ? behaves as if it were followed by ?: but named parentheses
2378 can still be used for capturing (and they will acquire numbers in the usual
2379 way).
2381 41. Redesigned the return codes from the match() function into yes/no/error so
2382 that errors can be passed back from deep inside the nested calls. A malloc
2383 failure while inside a recursive subpattern call now causes the
2384 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2386 42. It is now possible to set a limit on the number of times the match()
2387 function is called in a call to pcre_exec(). This facility makes it possible to
2388 limit the amount of recursion and backtracking, though not in a directly
2389 obvious way, because the match() function is used in a number of different
2390 circumstances. The count starts from zero for each position in the subject
2391 string (for non-anchored patterns). The default limit is, for compatibility, a
2392 large number, namely 10 000 000. You can change this in two ways:
2394 (a) When configuring PCRE before making, you can use --with-match-limit=n
2395 to set a default value for the compiled library.
2397 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2398 a different value is set. See 45 below.
2400 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2402 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2403 of things that can be changed at compile time. The first argument specifies
2404 what is wanted and the second points to where the information is to be placed.
2405 The current list of available information is:
2409 The output is an integer that is set to one if UTF-8 support is available;
2410 otherwise it is set to zero.
2414 The output is an integer that it set to the value of the code that is used for
2415 newline. It is either LF (10) or CR (13).
2419 The output is an integer that contains the number of bytes used for internal
2420 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2424 The output is an integer that contains the threshold above which the POSIX
2425 interface uses malloc() for output vectors. See item 31 above.
2429 The output is an unsigned integer that contains the default limit of the number
2430 of match() calls in a pcre_exec() execution. See 42 above.
2432 44. pcretest has been upgraded by the addition of the -C option. This causes it
2433 to extract all the available output from the new pcre_config() function, and to
2434 output it. The program then exits immediately.
2436 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2437 order to support additional features. One way would have been to define
2438 pcre_exec2() (for example) with extra arguments, but this would not have been
2439 extensible, and would also have required all calls to the original function to
2440 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2441 is used for passing in "extra" data from pcre_study().
2443 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2444 contains the following fields:
2446 flags a bitmap indicating which of the following fields are set
2447 study_data opaque data from pcre_study()
2448 match_limit a way of specifying a limit on match() calls for a specific
2449 call to pcre_exec()
2450 callout_data data for callouts (see 49 below)
2452 The flag bits are also defined in pcre.h, and are
2458 The pcre_study() function now returns one of these new pcre_extra blocks, with
2459 the actual study data pointed to by the study_data field, and the
2460 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2461 before. That is, this change is entirely upwards-compatible and requires no
2462 change to existing code.
2464 If you want to pass in additional data to pcre_exec(), you can either place it
2465 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2466 block.
2468 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2469 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2470 times with different match limits, until it finds the minimum value needed for
2471 pcre_exec() to complete. The value is then output. This can be instructive; for
2472 most simple matches the number is quite small, but for pathological cases it
2473 gets very large very quickly.
2475 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2476 returns the size of the data block pointed to by the study_data field in a
2477 pcre_extra block, that is, the value that was passed as the argument to
2478 pcre_malloc() when PCRE was getting memory in which to place the information
2479 created by pcre_study(). The fourth argument should point to a size_t variable.
2480 pcretest has been extended so that this information is shown after a successful
2481 pcre_study() call when information about the compiled regex is being displayed.
2483 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2484 because what follows is always an absolute path. (Later: it turns out that this
2485 is more than cosmetic for MinGW, because it doesn't like empty path
2486 components.)
2488 49. Some changes have been made to the callout feature (see 28 above):
2490 (i) A callout function now has three choices for what it returns:
2492 0 => success, carry on matching
2493 > 0 => failure at this point, but backtrack if possible
2494 < 0 => serious error, return this value from pcre_exec()
2496 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2497 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2498 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2499 use by callout functions. It will never be used by PCRE itself.
2501 (ii) The pcre_extra structure (see 45 above) has a void * field called
2502 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2503 pcre_callout_block structure has a field of the same name. The contents of
2504 the field passed in the pcre_extra structure are passed to the callout
2505 function in the corresponding field in the callout block. This makes it
2506 easier to use the same callout-containing regex from multiple threads. For
2507 testing, the pcretest program has a new data escape
2509 \C*n pass the number n (may be negative) as callout_data
2511 If the callout function in pcretest receives a non-zero value as
2512 callout_data, it returns that value.
2514 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2515 there were some redundant $(CFLAGS) in commands that are now specified as
2516 $(LINK), which already includes $(CFLAGS).
2518 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2519 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2520 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2521 one-byte characters throughout. Note that case-insensitive matching applies
2522 only to characters whose values are less than 256. PCRE doesn't support the
2523 notion of cases for higher-valued characters.
2525 (i) A character class whose characters are all within 0-255 is handled as
2526 a bit map, and the map is inverted for negative classes. Previously, a
2527 character > 255 always failed to match such a class; however it should
2528 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2530 (ii) A negated character class with a single character < 255 is coded as
2531 "not this character" (OP_NOT). This wasn't working properly when the test
2532 character was multibyte, either singly or repeated.
2534 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2535 mode, for example: \x{100}{2,3}.
2537 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2538 singly or repeated) now correctly test multibyte characters. However,
2539 PCRE doesn't recognize any characters with values greater than 255 as
2540 digits, spaces, or word characters. Such characters always match \D, \S,
2541 and \W, and never match \d, \s, or \w.
2543 (v) Classes may now contain characters and character ranges with values
2544 greater than 255. For example: [ab\x{100}-\x{400}].
2546 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2547 PCRE in UTF-8 mode.
2549 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2550 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2551 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2552 value.)
2554 53. The single man page has become too large. I have therefore split it up into
2555 a number of separate man pages. These also give rise to individual HTML pages;
2556 these are now put in a separate directory, and there is an index.html page that
2557 lists them all. Some hyperlinking between the pages has been installed.
2559 54. Added convenience functions for handling named capturing parentheses.
2561 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2562 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2563 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2564 are faulted.
2566 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2567 calling configure. These values are used when compiling the dftables.c program
2568 which is run to generate the source of the default character tables. They
2569 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2570 you will need to set these values.
2572 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2575 Version 3.9 02-Jan-02
2576 ---------------------
2578 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2580 2. If --disable-static was given, the building process failed when trying to
2581 build pcretest and pcregrep. (For some reason it was using libtool to compile
2582 them, which is not right, as they aren't part of the library.)
2585 Version 3.8 18-Dec-01
2586 ---------------------
2588 1. The experimental UTF-8 code was completely screwed up. It was packing the
2589 bytes in the wrong order. How dumb can you get?
2592 Version 3.7 29-Oct-01
2593 ---------------------
2595 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2596 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2597 this didn't happen under Solaris 8, where I normally test things.
2599 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2600 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2601 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2602 this fix an hour or so after the initial 3.7 release.)
2605 Version 3.6 23-Oct-01
2606 ---------------------
2608 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2609 offsets passed as NULL with zero offset count.
2611 2. The config.guess and config.sub files had not been updated when I moved to
2612 the latest autoconf.
2615 Version 3.5 15-Aug-01
2616 ---------------------
2618 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2619 had been forgotten.
2621 2. By using declared but undefined structures, we can avoid using "void"
2622 definitions in pcre.h while keeping the internal definitions of the structures
2623 private.
2625 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2626 user point of view, this means that both static and shared libraries are built
2627 by default, but this can be individually controlled. More of the work of
2628 handling this static/shared cases is now inside libtool instead of PCRE's make
2629 file.
2631 4. The pcretest utility is now installed along with pcregrep because it is
2632 useful for users (to test regexs) and by doing this, it automatically gets
2633 relinked by libtool. The documentation has been turned into a man page, so
2634 there are now .1, .txt, and .html versions in /doc.
2636 5. Upgrades to pcregrep:
2637 (i) Added long-form option names like gnu grep.
2638 (ii) Added --help to list all options with an explanatory phrase.
2639 (iii) Added -r, --recursive to recurse into sub-directories.
2640 (iv) Added -f, --file to read patterns from a file.
2642 6. pcre_exec() was referring to its "code" argument before testing that
2643 argument for NULL (and giving an error if it was NULL).
2645 7. Upgraded Makefile.in to allow for compiling in a different directory from
2646 the source directory.
2648 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2649 options bits, the pointer it was passed was to an int instead of to an unsigned
2650 long int. This mattered only on 64-bit systems.
2652 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2653 generated) instead of pcre.in, which it its source. Also made the same change
2654 in several of the .c files.
2656 10. A new release of gcc defines printf() as a macro, which broke pcretest
2657 because it had an ifdef in the middle of a string argument for printf(). Fixed
2658 by using separate calls to printf().
2660 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2661 script, to force use of CR or LF instead of \n in the source. On non-Unix
2662 systems, the value can be set in config.h.
2664 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2665 absolute limit. Changed the text of the error message to make this clear, and
2666 likewise updated the man page.
2668 13. The limit of 99 on the number of capturing subpatterns has been removed.
2669 The new limit is 65535, which I hope will not be a "real" limit.
2672 Version 3.4 22-Aug-00
2673 ---------------------
2675 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2677 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2680 Version 3.3 01-Aug-00
2681 ---------------------
2683 1. If an octal character was given, but the value was greater than \377, it
2684 was not getting masked to the least significant bits, as documented. This could
2685 lead to crashes in some systems.
2687 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2688 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2690 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2691 These just pass their arguments on to (pcre_free)(), but they are provided
2692 because some uses of PCRE bind it to non-C systems that can call its functions,
2693 but cannot call free() or pcre_free() directly.
2695 4. Add "make test" as a synonym for "make check". Corrected some comments in
2696 the Makefile.
2698 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2699 Makefile.
2701 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2702 command called pgrep for grepping around the active processes.
2704 7. Added the beginnings of support for UTF-8 character strings.
2706 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2707 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2708 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2709 out for the ar command.)
2712 Version 3.2 12-May-00
2713 ---------------------
2715 This is purely a bug fixing release.
2717 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2718 of ZA. This was just one example of several cases that could provoke this bug,
2719 which was introduced by change 9 of version 2.00. The code for breaking
2720 infinite loops after an iteration that matches an empty string was't working
2721 correctly.
2723 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2724 when matched against abbab (for example). After matching an empty string, it
2725 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2726 caused it to match further down the string than it should.
2728 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2729 was there because it doesn't seem to be needed, and it causes trouble on some
2730 systems, as it is not a Standard C header. It has been removed.
2732 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2733 were reported on the Macintosh. The changes were from
2735 while ((c = *(++ptr)) != 0 && c != '\n');
2736 to
2737 while ((c = *(++ptr)) != 0 && c != '\n') ;
2739 Totally extraordinary, but if that's what it takes...
2741 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2742 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2743 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2744 assumes the way PCRE uses memmove() (always moving upwards).
2746 6. PCRE is being used in one environment where strchr() is not available. There
2747 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2748 faster code anyway.
2751 Version 3.1 09-Feb-00
2752 ---------------------
2754 The only change in this release is the fixing of some bugs in Makefile.in for
2755 the "install" target:
2757 (1) It was failing to install pcreposix.h.
2759 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2762 Version 3.0 01-Feb-00
2763 ---------------------
2765 1. Add support for the /+ modifier to perltest (to output $` like it does in
2766 pcretest).
2768 2. Add support for the /g modifier to perltest.
2770 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2771 matches null strings.
2773 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2774 pattern. Perl treats empty patterns specially - it reuses the most recent
2775 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2776 effect.
2778 5. The POSIX interface was broken in that it was just handing over the POSIX
2779 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2780 required a bigger vector, with some working space on the end. This means that
2781 the POSIX wrapper now has to get and free some memory, and copy the results.
2783 6. Added some simple autoconf support, placing the test data and the
2784 documentation in separate directories, re-organizing some of the
2785 information files, and making it build pcre-config (a GNU standard). Also added
2786 libtool support for building PCRE as a shared library, which is now the
2787 default.
2789 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2790 09 are not valid octal constants. Single digits will be used for minor values
2791 less than 10.
2793 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2794 existing programs that set these in the POSIX interface can use PCRE without
2795 modification.
2797 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2798 return all that pcre_info() returns, plus additional data. The pcre_info()
2799 function is retained for compatibility, but is considered to be obsolete.
2801 10. Added experimental recursion feature (?R) to handle one common case that
2802 Perl 5.6 will be able to do with (?p{...}).
2804 11. Added support for POSIX character classes like [:alpha:], which Perl is
2805 adopting.
2808 Version 2.08 31-Aug-99
2809 ----------------------
2811 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2812 trying to match at the startoffset position, but instead was moving forward to
2813 the next newline as if a previous match had failed.
2815 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2816 and could get into a loop if a null string was matched other than at the start
2817 of the subject.
2819 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2820 be distinguished at compile time, and for completeness also added PCRE_DATE.
2822 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2823 in GnuWin32 environments.
2826 Version 2.07 29-Jul-99
2827 ----------------------
2829 1. The documentation is now supplied in plain text form and HTML as well as in
2830 the form of man page sources.
2832 2. C++ compilers don't like assigning (void *) values to other pointer types.
2833 In particular this affects malloc(). Although there is no problem in Standard
2834 C, I've put in casts to keep C++ compilers happy.
2836 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2837 should be (const char *).
2839 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2840 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2841 However, I haven't made this a standard facility. The documentation doesn't
2842 mention it, and the Makefile doesn't support it.
2844 5. The Makefile now contains an "install" target, with editable destinations at
2845 the top of the file. The pcretest program is not installed.
2847 6. pgrep -V now gives the PCRE version number and date.
2849 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2850 causing the entire string to be ignored, instead of just the last character.
2852 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2853 non-matching string, it can take a very, very long time, even for strings of
2854 quite modest length, because of the nested recursion. PCRE now does better in
2855 some of these cases. It does this by remembering the last required literal
2856 character in the pattern, and pre-searching the subject to ensure it is present
2857 before running the real match. In other words, it applies a heuristic to detect
2858 some types of certain failure quickly, and in the above example, if presented
2859 with a string that has no trailing " it gives "no match" very quickly.
2861 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2862 other alternatives are tried instead.
2865 Version 2.06 09-Jun-99
2866 ----------------------
2868 1. Change pcretest's output for amount of store used to show just the code
2869 space, because the remainder (the data block) varies in size between 32-bit and
2870 64-bit systems.
2872 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2873 start matching at. This allows lookbehinds to work when searching for multiple
2874 occurrences in a string.
2876 3. Added additional options to pcretest for testing multiple occurrences:
2878 /+ outputs the rest of the string that follows a match
2879 /g loops for multiple occurrences, using the new startoffset argument
2880 /G loops for multiple occurrences by passing an incremented pointer
2882 4. PCRE wasn't doing the "first character" optimization for patterns starting
2883 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2884 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2885 the letter 'x'. On long subject strings, this gives a significant speed-up.
2888 Version 2.05 21-Apr-99
2889 ----------------------
2891 1. Changed the type of magic_number from int to long int so that it works
2892 properly on 16-bit systems.
2894 2. Fixed a bug which caused patterns starting with .* not to work correctly
2895 when the subject string contained newline characters. PCRE was assuming
2896 anchoring for such patterns in all cases, which is not correct because .* will
2897 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2898 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2899 must be retried after every newline in the subject.
2902 Version 2.04 18-Feb-99
2903 ----------------------
2905 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2906 computation of the store needed to hold the pattern was incorrect (too large).
2907 If such patterns were nested a few deep, this could multiply and become a real
2908 problem.
2910 2. Added /M option to pcretest to show the memory requirement of a specific
2911 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2913 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2914 compiled in such a way that the backtracking after subsequent failure was
2915 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2916 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2919 Version 2.03 02-Feb-99
2920 ----------------------
2922 1. Fixed typo and small mistake in man page.
2924 2. Added 4th condition (GPL supersedes if conflict) and created separate
2925 LICENCE file containing the conditions.
2927 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2928 Perl, that is the internal \ allows the delimiter to be included in the
2929 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2930 the final delimiter, add \ to the end of the pattern (to test the error).
2932 4. Added the convenience functions for extracting substrings after a successful
2933 match. Updated pcretest to make it able to test these functions.
2936 Version 2.02 14-Jan-99
2937 ----------------------
2939 1. Initialized the working variables associated with each extraction so that
2940 their saving and restoring doesn't refer to uninitialized store.
2942 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2943 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2944 fix the problem.
2946 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2947 calls, and wasn't printing the correct value for compiling calls. Increased the
2948 default value of LOOPREPEAT, and the number of significant figures in the
2949 times.
2951 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2953 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2954 a building problem on Windows NT with a FAT file system.
2957 Version 2.01 21-Oct-98
2958 ----------------------
2960 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2961 to character tables built by pcre_maketables() in the current locale. If NULL
2962 is passed, the default tables are used.
2965 Version 2.00 24-Sep-98
2966 ----------------------
2968 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2969 it any more.
2971 2. Allow quantification of (?>) groups, and make it work correctly.
2973 3. The first character computation wasn't working for (?>) groups.
2975 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2976 end of the subject) and add 5.005's \z, which really does match only at the
2977 very end of the subject.
2979 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2981 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2982 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2983 localized options. All options to pcre_study() were also removed.
2985 7. Add other new features from 5.005:
2987 $(?<= positive lookbehind
2988 $(?<! negative lookbehind
2989 (?imsx-imsx) added the unsetting capability
2990 such a setting is global if at outer level; local otherwise
2991 (?imsx-imsx:) non-capturing groups with option setting
2992 (?(cond)re|re) conditional pattern matching
2994 A backreference to itself in a repeated group matches the previous
2995 captured string.
2997 8. General tidying up of studying (both automatic and via "study")
2998 consequential on the addition of new assertions.
3000 9. As in 5.005, unlimited repeated groups that could match an empty substring
3001 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3002 runtime if any iteration does actually match an empty substring.
3004 10. Include the RunTest script in the distribution.
3006 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3007 discrepancies, some of which were old and were also with respect to 5.004. They
3008 have now been fixed.
3011 Version 1.09 28-Apr-98
3012 ----------------------
3014 1. A negated single character class followed by a quantifier with a minimum
3015 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3016 program crashes, or just wrong answers. This did not apply to negated classes
3017 containing more than one character, or to minima other than one.
3020 Version 1.08 27-Mar-98
3021 ----------------------
3023 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3025 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3026 latter must appear before anything that relies on it in the pattern.
3029 Version 1.07 16-Feb-98
3030 ----------------------
3032 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3033 repeat of a potentially empty string).
3036 Version 1.06 23-Jan-98
3037 ----------------------
3039 1. Added Markus Oberhumer's little patches for C++.
3041 2. Literal strings longer than 255 characters were broken.
3044 Version 1.05 23-Dec-97
3045 ----------------------
3047 1. Negated character classes containing more than one character were failing if
3048 PCRE_CASELESS was set at run time.
3051 Version 1.04 19-Dec-97
3052 ----------------------
3054 1. Corrected the man page, where some "const" qualifiers had been omitted.
3056 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3057 input syntax.
3059 3. Fixed memory leak which occurred when a regex with back references was
3060 matched with an offsets vector that wasn't big enough. The temporary memory
3061 that is used in this case wasn't being freed if the match failed.
3063 4. Tidied pcretest to ensure it frees memory that it gets.
3065 5. Temporary memory was being obtained in the case where the passed offsets
3066 vector was exactly big enough.
3068 6. Corrected definition of offsetof() from change 5 below.
3070 7. I had screwed up change 6 below and broken the rules for the use of
3071 setjmp(). Now fixed.
3074 Version 1.03 18-Dec-97
3075 ----------------------
3077 1. A erroneous regex with a missing opening parenthesis was correctly
3078 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3079 on some systems.
3081 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3082 it was reported that one broken compiler failed on the former because "code" is
3083 also an independent variable.
3085 3. The erroneous regex a[]b caused an array overrun reference.
3087 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3088 fail on data ending with that character. (It was going on too far, and checking
3089 the next character, typically a binary zero.) This was specific to the
3090 optimized code for single-character negative classes.
3092 5. Added a contributed patch from the TIN world which does the following:
3094 + Add an undef for memmove, in case the the system defines a macro for it.
3096 + Add a definition of offsetof(), in case there isn't one. (I don't know
3097 the reason behind this - offsetof() is part of the ANSI standard - but
3098 it does no harm).
3100 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3101 most of the places where whitespace preceded '#'. I have given up and
3102 allowed the remaining 2 cases to be at the margin.
3104 + Rename some variables in pcre to eliminate shadowing. This seems very
3105 pedantic, but does no harm, of course.
3107 6. Moved the call to setjmp() into its own function, to get rid of warnings
3108 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3110 7. Constructs such as \d{8,} were compiling into the equivalent of
3111 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3112 outcome, but in this particular case used more store than had been allocated,
3113 which caused the bug to be discovered because it threw up an internal error.
3115 8. The debugging code in both pcre and pcretest for outputting the compiled
3116 form of a regex was going wrong in the case of back references followed by
3117 curly-bracketed repeats.
3120 Version 1.02 12-Dec-97
3121 ----------------------
3123 1. Typos in pcre.3 and comments in the source fixed.
3125 2. Applied a contributed patch to get rid of places where it used to remove
3126 'const' from variables, and fixed some signed/unsigned and uninitialized
3127 variable warnings.
3129 3. Added the "runtest" target to Makefile.
3131 4. Set default compiler flag to -O2 rather than just -O.
3134 Version 1.01 19-Nov-97
3135 ----------------------
3137 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3138 like /([ab]*)*/, that is, for classes with more than one character in them.
3140 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3141 as /((?>a*))*/ (a PCRE_EXTRA facility).
3144 Version 1.00 18-Nov-97
3145 ----------------------
3147 1. Added compile-time macros to support systems such as SunOS4 which don't have
3148 memmove() or strerror() but have other things that can be used instead.
3150 2. Arranged that "make clean" removes the executables.
3153 Version 0.99 27-Oct-97
3154 ----------------------
3156 1. Fixed bug in code for optimizing classes with only one character. It was
3157 initializing a 32-byte map regardless, which could cause it to run off the end
3158 of the memory it had got.
3160 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3163 Version 0.98 22-Oct-97
3164 ----------------------
3166 1. Fixed bug in code for handling temporary memory usage when there are more
3167 back references than supplied space in the ovector. This could cause segfaults.
3170 Version 0.97 21-Oct-97
3171 ----------------------
3173 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3175 2. Optimized negated single characters not to use a bit map.
3177 3. Brought error texts together as macro definitions; clarified some of them;
3178 fixed one that was wrong - it said "range out of order" when it meant "invalid
3179 escape sequence".
3181 4. Changed some char * arguments to const char *.
3183 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3185 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3186 pcretest.
3189 Version 0.96 16-Oct-97
3190 ----------------------
3192 1. Added a simple "pgrep" utility to the distribution.
3194 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3195 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3196 where "ddd" means "one or more decimal digits".
3198 3. Fixed serious bug. If a pattern had a back reference, but the call to
3199 pcre_exec() didn't supply a large enough ovector to record the related
3200 identifying subpattern, the match always failed. PCRE now remembers the number
3201 of the largest back reference, and gets some temporary memory in which to save
3202 the offsets during matching if necessary, in order to ensure that
3203 backreferences always work.
3205 4. Increased the compatibility with Perl in a number of ways:
3207 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3208 to request this handling. The option can be set at compile or exec time.
3210 (b) $ matches before a terminating newline by default; an option
3211 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3212 mode). The option can be set at compile or exec time.
3214 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3215 the same as Perl's. If the decimal number it represents is less than 10
3216 or there aren't that many previous left capturing parentheses, an octal
3217 escape is read. Inside a character class, it's always an octal escape,
3218 even if it is a single digit.
3220 (d) An escaped but undefined alphabetic character is taken as a literal,
3221 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3222 escapes.
3224 (e) {0} is now permitted. (The previous item is removed from the compiled
3225 pattern).
3227 5. Changed all the names of code files so that the basic parts are no longer
3228 than 10 characters, and abolished the teeny "globals.c" file.
3230 6. Changed the handling of character classes; they are now done with a 32-byte
3231 bit map always.
3233 7. Added the -d and /D options to pcretest to make it possible to look at the
3234 internals of compilation without having to recompile pcre.
3237 Version 0.95 23-Sep-97
3238 ----------------------
3240 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3241 \x20 at the start of a run of normal characters. These were being treated as
3242 real characters, instead of the source characters being re-checked.
3245 Version 0.94 18-Sep-97
3246 ----------------------
3248 1. The functions are now thread-safe, with the caveat that the global variables
3249 containing pointers to malloc() and free() or alternative functions are the
3250 same for all threads.
3252 2. Get pcre_study() to generate a bitmap of initial characters for non-
3253 anchored patterns when this is possible, and use it if passed to pcre_exec().
3256 Version 0.93 15-Sep-97
3257 ----------------------
3259 1. /(b)|(:+)/ was computing an incorrect first character.
3261 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3262 but not actually doing anything yet.
3264 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3265 as Perl does (e.g. [-az] or [az-]).
3267 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3268 all possible positions.
3270 5. Split up into different modules to avoid including unneeded functions in a
3271 compiled binary. However, compile and exec are still in one module. The "study"
3272 function is split off.
3274 6. The character tables are now in a separate module whose source is generated
3275 by an auxiliary program - but can then be edited by hand if required. There are
3276 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3277 toupper() in the code.
3279 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3280 make them global. Abolish the function for setting them, as the caller can now
3281 set them directly.
3284 Version 0.92 11-Sep-97
3285 ----------------------
3287 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3288 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3290 2. Caseless matching was not working in character classes if the characters in
3291 the pattern were in upper case.
3293 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3295 4. Make PCRE_ANCHORED public and accept as a compile option.
3297 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3298 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3299 pass them.
3301 6. Give an error if bad option bits passed at compile or run time.
3303 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3304 pcretest to cause it to pass that flag.
3306 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3307 options, and the first character, if set.
3309 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3312 Version 0.91 10-Sep-97
3313 ----------------------
3315 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3316 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3318 2. PCRE was looping on encountering an indefinitely repeated back reference to
3319 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3320 Perl does - treats the match as successful.
3322 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5