7 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
8 a string to a long long, pretend we don't even have a long long." This is
9 done by checking for the strtoq, strtoll, and _strtoi64 functions.
11 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
12 pre-7.6 versions, which defined a global no_arg variable instead of putting
13 it in the RE class. (See also #8 below.)
15 3. Remove a line of dead code, identified by coverity and reported by Nuno
18 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
20 (1) The include/exclude patterns were being applied to the whole pathnames
21 of files, instead of just to the final components.
23 (2) If there was more than one level of directory, the subdirectories were
24 skipped unless they satisfied the include/exclude conditions. This is
25 inconsistent with GNU grep (and could even be seen as contrary to the
26 pcregrep specification - which I improved to make it absolutely clear).
27 The action now is always to scan all levels of directory, and just
28 apply the include/exclude patterns to regular files.
30 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
31 --exclude_dir in the tests to avoid scanning .svn directories.
33 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
34 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
35 doesn't support NULs in patterns.
37 7. Added some missing "const"s to declarations of static tables in
38 pcre_compile.c and pcre_dfa_exec.c.
40 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
41 caused by fix #2 above. (Subsequently also a second patch to fix the
42 first patch. And a third patch - this was a messy problem.)
44 9. Applied Craig's patch to remove the use of push_back().
46 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
47 matching function regexec().
49 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
50 which, however, unlike Perl's \g{...}, are subroutine calls, not back
51 references. PCRE supports relative numbers with this syntax (I don't think
54 12. Previously, a group with a zero repeat such as (...){0} was completely
55 omitted from the compiled regex. However, this means that if the group
56 was called as a subroutine from elsewhere in the pattern, things went wrong
57 (an internal error was given). Such groups are now left in the compiled
58 pattern, with a new opcode that causes them to be skipped at execution
61 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
62 to the way PCRE behaves:
64 (a) A lone ] character is dis-allowed (Perl treats it as data).
66 (b) A back reference to an unmatched subpattern matches an empty string
67 (Perl fails the current match path).
69 (c) A data ] in a character class must be notated as \] because if the
70 first data character in a class is ], it defines an empty class. (In
71 Perl it is not possible to have an empty class.) The empty class []
72 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
73 The negative empty class [^] matches any one character, independently
74 of the DOTALL setting.
76 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
77 non-existent subpattern following a character class starting with ']' and
78 containing () gave an internal compiling error instead of "reference to
79 non-existent subpattern". Fortunately, when the pattern did exist, the
80 compiled code was correct. (When scanning forwards to check for the
81 existencd of the subpattern, it was treating the data ']' as terminating
82 the class, so got the count wrong. When actually compiling, the reference
83 was subsequently set up correctly.)
85 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
86 it was being rejected as not supported by pcre_dfa_exec(), even though
87 other assertions are supported. I have made pcre_dfa_exec() support
90 16. The implementation of 13c above involved the invention of a new opcode,
91 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
92 cannot be changed at match time, I realized I could make a small
93 improvement to matching performance by compiling OP_ALLANY instead of
94 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
97 17. Compiling pcretest on Windows with readline support failed without the
98 following two fixes: (1) Make the unistd.h include conditional on
99 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
101 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
102 ncurses library to be included for pcretest when ReadLine support is
103 requested, but also to allow for it to be overridden. This patch came from
106 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
107 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
108 any errors with the current Unicode tables. Thanks to Peter Kankowski for
112 Version 7.6 28-Jan-08
113 ---------------------
115 1. A character class containing a very large number of characters with
116 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
119 2. Patch to cut out the "long long" test in pcrecpp_unittest when
120 HAVE_LONG_LONG is not defined.
122 3. Applied Christian Ehrlicher's patch to update the CMake build files to
123 bring them up to date and include new features. This patch includes:
125 - Fixed PH's badly added libz and libbz2 support.
126 - Fixed a problem with static linking.
127 - Added pcredemo. [But later removed - see 7 below.]
128 - Fixed dftables problem and added an option.
129 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
131 - Added readline support for pcretest.
132 - Added an listing of the option settings after cmake has run.
134 4. A user submitted a patch to Makefile that makes it easy to create
135 "pcre.dll" under mingw when using Configure/Make. I added stuff to
136 Makefile.am that cause it to include this special target, without
137 affecting anything else. Note that the same mingw target plus all
138 the other distribution libraries and programs are now supported
139 when configuring with CMake (see 6 below) instead of with
142 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
143 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
144 exported in the Windows port". It has not yet been confirmed that the patch
145 solves the problem, but it does no harm.
147 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
148 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
149 with CMake, and also correct the comment about stack recursion.
151 7. Remove the automatic building of pcredemo from the ./configure system and
152 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
153 of a program that users should build themselves after PCRE is installed, so
154 building it automatically is not really right. What is more, it gave
155 trouble in some build environments.
157 8. Further tidies to CMakeLists.txt from Sheri and Christian.
160 Version 7.5 10-Jan-08
161 ---------------------
163 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
164 values in parens when parsing an RE using the C++ wrapper."
166 2. Negative specials like \S did not work in character classes in UTF-8 mode.
167 Characters greater than 255 were excluded from the class instead of being
170 3. The same bug as (2) above applied to negated POSIX classes such as
173 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
174 defined or documented. It seems to have been a typo for PCRE_STATIC, so
177 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
178 first named subpattern) and a construct such as (?&a) would reference the
179 first named subpattern whose name started with "a" (in other words, the
180 length check was missing). Both these problems are fixed. "Subpattern name
181 expected" is now given for (?&) (a zero-length name), and this patch also
182 makes it give the same error for \k'' (previously it complained that that
183 was a reference to a non-existent subpattern).
185 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
186 this is right because (?- can be followed by option settings as well as by
187 digits. I have, however, made the messages clearer.
189 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
190 than the number used in the conditional) now cause a compile-time error.
191 This is actually not compatible with Perl, which accepts such patterns, but
192 treats the conditional as always being FALSE (as PCRE used to), but it
193 seems to me that giving a diagnostic is better.
195 8. Change "alphameric" to the more common word "alphanumeric" in comments
198 9. Fix two occurrences of "backslash" in comments that should have been
201 10. Remove two redundant lines of code that can never be obeyed (their function
202 was moved elsewhere).
204 11. The program that makes PCRE's Unicode character property table had a bug
205 which caused it to generate incorrect table entries for sequences of
206 characters that have the same character type, but are in different scripts.
207 It amalgamated them into a single range, with the script of the first of
208 them. In other words, some characters were in the wrong script. There were
209 thirteen such cases, affecting characters in the following ranges:
225 12. The -o option (show only the matching part of a line) for pcregrep was not
226 compatible with GNU grep in that, if there was more than one match in a
227 line, it showed only the first of them. It now behaves in the same way as
230 13. If the -o and -v options were combined for pcregrep, it printed a blank
231 line for every non-matching line. GNU grep prints nothing, and pcregrep now
232 does the same. The return code can be used to tell if there were any
235 14. Added --file-offsets and --line-offsets to pcregrep.
237 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
238 infinitely looping recursion. The bug was that positive lookaheads were not
239 being skipped when checking for a possible empty match (negative lookaheads
240 and both kinds of lookbehind were skipped).
242 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
243 inclusion of <windows.h> to before rather than after the definition of
244 INVALID_FILE_ATTRIBUTES (patch from David Byron).
246 17. Specifying a possessive quantifier with a specific limit for a Unicode
247 character property caused pcre_compile() to compile bad code, which led at
248 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
249 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
250 caused the error; without that there was no problem.
252 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
254 19. Added --enable-pcretest-libreadline.
256 20. In pcrecpp.cc, the variable 'count' was incremented twice in
257 RE::GlobalReplace(). As a result, the number of replacements returned was
258 double what it should be. I removed one of the increments, but Craig sent a
259 later patch that removed the other one (the right fix) and added unit tests
260 that check the return values (which was not done before).
262 21. Several CMake things:
264 (1) Arranged that, when cmake is used on Unix, the libraries end up with
265 the names libpcre and libpcreposix, not just pcre and pcreposix.
267 (2) The above change means that pcretest and pcregrep are now correctly
268 linked with the newly-built libraries, not previously installed ones.
270 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
272 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
273 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
274 UTF-8 newline character). The key issue is that the pattern starts .*;
275 this means that the match must be either at the beginning, or after a
276 newline. The bug was in the code for advancing after a failed match and
277 checking that the new position followed a newline. It was not taking
278 account of UTF-8 characters correctly.
280 23. PCRE was behaving differently from Perl in the way it recognized POSIX
281 character classes. PCRE was not treating the sequence [:...:] as a
282 character class unless the ... were all letters. Perl, however, seems to
283 allow any characters between [: and :], though of course it rejects as
284 unknown any "names" that contain non-letters, because all the known class
285 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
286 for example, whereas PCRE did not - it did not recognize a POSIX character
287 class. This seemed a bit dangerous, so the code has been changed to be
288 closer to Perl. The behaviour is not identical to Perl, because PCRE will
289 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
290 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
291 Perl does, and where it didn't before.
293 24. Rewrite so as to remove the single use of %n from pcregrep because in some
294 Windows environments %n is disabled by default.
297 Version 7.4 21-Sep-07
298 ---------------------
300 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
301 means that a class such as [\s] counted as "explicit reference to CR or
302 LF". That isn't really right - the whole point of the change was to try to
303 help when there was an actual mention of one of the two characters. So now
304 the change happens only if \r or \n (or a literal CR or LF) character is
307 2. The 32-bit options word was also used for 6 internal flags, but the numbers
308 of both had grown to the point where there were only 3 bits left.
309 Fortunately, there was spare space in the data structure, and so I have
310 moved the internal flags into a new 16-bit field to free up more option
313 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
314 but did not set the internal JCHANGED flag - either of these is enough to
315 control the way the "get" function works - but the PCRE_INFO_JCHANGED
316 facility is supposed to tell if (?J) was ever used, so now (?J) at the
317 start sets both bits.
319 4. Added options (at build time, compile time, exec time) to change \R from
320 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
322 5. doc/pcresyntax.html was missing from the distribution.
324 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
325 compatibility, even though it is no longer used.
327 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
328 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
329 windows.h file is present (where different names are used). [This was
330 reversed later after testing - see 16 below.]
332 8. Changed all #include <config.h> to #include "config.h". There were also
333 some further <pcre.h> cases that I changed to "pcre.h".
335 9. When pcregrep was used with the --colour option, it missed the line ending
336 sequence off the lines that it output.
338 10. It was pointed out to me that arrays of string pointers cause lots of
339 relocations when a shared library is dynamically loaded. A technique of
340 using a single long string with a table of offsets can drastically reduce
341 these. I have refactored PCRE in four places to do this. The result is
345 After changing UCP table: 187
346 After changing error message table: 43
347 After changing table of "verbs" 36
348 After changing table of Posix names 22
350 Thanks to the folks working on Gregex for glib for this insight.
352 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
353 unicode-properties was also set.
355 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
357 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
358 checked only for CRLF.
360 14. Added casts to pcretest.c to avoid compiler warnings.
362 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
364 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
365 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
366 entirely. This removes changes made in 7 above.
368 17. The CMake files have been updated, and there is now more information about
369 building with CMake in the NON-UNIX-USE document.
372 Version 7.3 28-Aug-07
373 ---------------------
375 1. In the rejigging of the build system that eventually resulted in 7.1, the
376 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
377 brackets there is not right, since it causes compilers to look for an
378 installed pcre.h, not the version that is in the source that is being
379 compiled (which of course may be different). I have changed it back to:
383 I have a vague recollection that the change was concerned with compiling in
384 different directories, but in the new build system, that is taken care of
385 by the VPATH setting the Makefile.
387 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
388 when the subject happened to end in the byte 0x85 (e.g. if the last
389 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
390 characters but of course it shouldn't be taken as a newline when it is part
391 of another character. The bug was that, for an unlimited repeat of . in
392 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
393 characters when looking for a newline.
395 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
397 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
400 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
401 long printing in the pcrecpp unittest when running under MinGW.
403 6. ESC_K was left out of the EBCDIC table.
405 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
406 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
407 limit also applies to "virtual nesting" when a pattern is recursive, and in
408 this case 1000 isn't so big. I have been able to remove this limit at the
409 expense of backing off one optimization in certain circumstances. Normally,
410 when pcre_exec() would call its internal match() function recursively and
411 immediately return the result unconditionally, it uses a "tail recursion"
412 feature to save stack. However, when a subpattern that can match an empty
413 string has an unlimited repetition quantifier, it no longer makes this
414 optimization. That gives it a stack frame in which to save the data for
415 checking that an empty string has been matched. Previously this was taken
416 from the 1000-entry workspace that had been reserved. So now there is no
417 explicit limit, but more stack is used.
419 8. Applied Daniel's patches to solve problems with the import/export magic
420 syntax that is required for Windows, and which was going wrong for the
421 pcreposix and pcrecpp parts of the library. These were overlooked when this
422 problem was solved for the main library.
424 9. There were some crude static tests to avoid integer overflow when computing
425 the size of patterns that contain repeated groups with explicit upper
426 limits. As the maximum quantifier is 65535, the maximum group length was
427 set at 30,000 so that the product of these two numbers did not overflow a
428 32-bit integer. However, it turns out that people want to use groups that
429 are longer than 30,000 bytes (though not repeat them that many times).
430 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
431 made it possible to implement the integer overflow checks in a much more
432 dynamic way, which I have now done. The artificial limitation on group
433 length has been removed - we now have only the limit on the total length of
434 the compiled pattern, which depends on the LINK_SIZE setting.
436 10. Fixed a bug in the documentation for get/copy named substring when
437 duplicate names are permitted. If none of the named substrings are set, the
438 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
441 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
442 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
443 because the ] is interpreted as the first data character and the
444 terminating ] is not found. PCRE has been made compatible with Perl in this
445 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
446 cause memory overwriting.
448 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
449 string has been matched (to stop an infinite loop). It was not recognizing
450 a conditional subpattern that could match an empty string if that
451 subpattern was within another subpattern. For example, it looped when
452 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
453 condition was not nested. This bug has been fixed.
455 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
456 past the start of the subject in the presence of bytes with the top bit
457 set, for example "\x8aBCD".
459 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
460 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
462 14. Optimized (?!) to (*FAIL).
464 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
465 This restricts code points to be within the range 0 to 0x10FFFF, excluding
466 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
467 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
468 does: it's just the validity check that is more restrictive.
470 16. Inserted checks for integer overflows during escape sequence (backslash)
471 processing, and also fixed erroneous offset values for syntax errors during
472 backslash processing.
474 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
475 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
477 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
480 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
481 something other than just ASCII characters) inside a group that had an
482 unlimited repeat caused a loop at compile time (while checking to see
483 whether the group could match an empty string).
485 20. Debugging a pattern containing \p or \P could cause a crash. For example,
486 [\P{Any}] did so. (Error in the code for printing property names.)
488 21. An orphan \E inside a character class could cause a crash.
490 22. A repeated capturing bracket such as (A)? could cause a wild memory
491 reference during compilation.
493 23. There are several functions in pcre_compile() that scan along a compiled
494 expression for various reasons (e.g. to see if it's fixed length for look
495 behind). There were bugs in these functions when a repeated \p or \P was
496 present in the pattern. These operators have additional parameters compared
497 with \d, etc, and these were not being taken into account when moving along
498 the compiled data. Specifically:
500 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
503 (b) An item such as \pL+ within a repeated group could cause crashes or
506 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
507 "reference to non-existent subpattern" error.
509 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
511 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
512 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
514 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
516 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
517 character were causing crashes (broken optimization).
519 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
520 \p or \P) caused a compile-time loop.
522 28. More problems have arisen in unanchored patterns when CRLF is a valid line
523 break. For example, the unstudied pattern [\r\n]A does not match the string
524 "\r\nA" because change 7.0/46 below moves the current point on by two
525 characters after failing to match at the start. However, the pattern \nA
526 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
527 the same is true. There doesn't seem any very clean way out of this, but
528 what I have chosen to do makes the common cases work: PCRE now takes note
529 of whether there can be an explicit match for \r or \n anywhere in the
530 pattern, and if so, 7.0/46 no longer applies. As part of this change,
531 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
532 pattern has explicit CR or LF references.
534 29. Added (*CR) etc for changing newline setting at start of pattern.
537 Version 7.2 19-Jun-07
538 ---------------------
540 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
541 which is apparently normally available under Windows.
543 2. Re-jig the pcregrep tests with different newline settings in an attempt
544 to make them independent of the local environment's newline setting.
546 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
548 4. Some of the "internals" tests were previously cut out when the link size
549 was not 2, because the output contained actual offsets. The recent new
550 "Z" feature of pcretest means that these can be cut out, making the tests
551 usable with all link sizes.
553 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
554 stack recursion. This gives a massive performance boost under BSD, but just
555 a small improvement under Linux. However, it saves one field in the frame
558 6. Added more features from the forthcoming Perl 5.10:
560 (a) (?-n) (where n is a string of digits) is a relative subroutine or
561 recursion call. It refers to the nth most recently opened parentheses.
563 (b) (?+n) is also a relative subroutine call; it refers to the nth next
564 to be opened parentheses.
566 (c) Conditions that refer to capturing parentheses can be specified
567 relatively, for example, (?(-2)... or (?(+3)...
569 (d) \K resets the start of the current match so that everything before
572 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
574 (f) \g{name} is another synonym - part of Perl 5.10's unification of
577 (g) (?| introduces a group in which the numbering of parentheses in each
578 alternative starts with the same number.
580 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
582 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
585 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
586 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
587 for detecting groups that can match an empty string.
589 9. A pattern with a very large number of alternatives (more than several
590 hundred) was running out of internal workspace during the pre-compile
591 phase, where pcre_compile() figures out how much memory will be needed. A
592 bit of new cunning has reduced the workspace needed for groups with
593 alternatives. The 1000-alternative test pattern now uses 12 bytes of
594 workspace instead of running out of the 4096 that are available.
596 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
598 11. Applied patch from Google to remove an optimization that didn't quite work.
599 The report of the bug said:
601 pcrecpp::RE("a*").FullMatch("aaa") matches, while
602 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
603 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
605 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
606 it matched the wrong number of bytes.
609 Version 7.1 24-Apr-07
610 ---------------------
612 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
613 that is more "standard", making use of automake and other Autotools. There
614 is some re-arrangement of the files and adjustment of comments consequent
617 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
618 for recursive directory scanning broke on some systems because the files
619 are not scanned in any specific order and on different systems the order
620 was different. A call to "sort" has been inserted into RunGrepTest for the
621 approprate test as a short-term fix. In the longer term there may be an
624 3. I had an email from Eric Raymond about problems translating some of PCRE's
625 man pages to HTML (despite the fact that I distribute HTML pages, some
626 people do their own conversions for various reasons). The problems
627 concerned the use of low-level troff macros .br and .in. I have therefore
628 removed all such uses from the man pages (some were redundant, some could
629 be replaced by .nf/.fi pairs). The 132html script that I use to generate
630 HTML has been updated to handle .nf/.fi and to complain if it encounters
633 4. Updated comments in configure.ac that get placed in config.h.in and also
634 arranged for config.h to be included in the distribution, with the name
635 config.h.generic, for the benefit of those who have to compile without
636 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
638 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
639 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
640 makevp.bat for the current PCRE, using the additional files
641 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
643 6. A Windows user reported a minor discrepancy with test 2, which turned out
644 to be caused by a trailing space on an input line that had got lost in his
645 copy. The trailing space was an accident, so I've just removed it.
647 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
650 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
651 as "const" (a) because they are and (b) because it helps the PHP
652 maintainers who have recently made a script to detect big data structures
653 in the php code that should be moved to the .rodata section. I remembered
654 to update Builducptable as well, so it won't revert if ucptable.h is ever
657 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
658 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
659 order to be able to cut out the UTF-8 tables in the latter when UTF-8
660 support is not required. This saves 1.5-2K of code, which is important in
663 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
664 so as not to refer to the tables, even though these functions will never be
665 called when UTF-8 support is disabled. Otherwise there are problems with a
668 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
670 (a) It was defining its arguments as char * instead of void *.
672 (b) It was assuming that all moves were upwards in memory; this was true
673 a long time ago when I wrote it, but is no longer the case.
675 The emulated memove() is provided for those environments that have neither
676 memmove() nor bcopy(). I didn't think anyone used it these days, but that
677 is clearly not the case, as these two bugs were recently reported.
679 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
680 and Detrail to create the HTML documentation, the .txt form of the man
681 pages, and it removes trailing spaces from listed files. It also creates
682 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
683 case, it wraps all the #defines with #ifndefs. This script should be run
686 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
687 with Unicode property support.
689 (a) For a maximizing quantifier, if the two different cases of the
690 character were of different lengths in their UTF-8 codings (there are
691 some cases like this - I found 11), and the matching function had to
692 back up over a mixture of the two cases, it incorrectly assumed they
693 were both the same length.
695 (b) When PCRE was configured to use the heap rather than the stack for
696 recursion during matching, it was not correctly preserving the data for
697 the other case of a UTF-8 character when checking ahead for a match
698 while processing a minimizing repeat. If the check also involved
699 matching a wide character, but failed, corruption could cause an
700 erroneous result when trying to check for a repeat of the original
703 13. Some tidying changes to the testing mechanism:
705 (a) The RunTest script now detects the internal link size and whether there
706 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
707 values substituted by "configure". (The RunGrepTest script already did
708 this for UTF-8.) The configure.ac script no longer substitutes the
711 (b) The debugging options /B and /D in pcretest show the compiled bytecode
712 with length and offset values. This means that the output is different
713 for different internal link sizes. Test 2 is skipped for link sizes
714 other than 2 because of this, bypassing the problem. Unfortunately,
715 there was also a test in test 3 (the locale tests) that used /B and
716 failed for link sizes other than 2. Rather than cut the whole test out,
717 I have added a new /Z option to pcretest that replaces the length and
718 offset values with spaces. This is now used to make test 3 independent
719 of link size. (Test 2 will be tidied up later.)
721 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
722 segmentation fault instead of returning the appropriate error message.
724 15. In multiline mode when the newline sequence was set to "any", the pattern
725 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
726 This doesn't seem right; it now treats the CRLF combination as the line
727 ending, and so does not match in that case. It's only a pattern such as ^$
728 that would hit this one: something like ^ABC$ would have failed after \r
729 and then tried again after \r\n.
731 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
732 in an attempt to make files that differ only in their line terminators
733 compare equal. This works on Linux.
735 17. Under certain error circumstances pcregrep might try to free random memory
736 as it exited. This is now fixed, thanks to valgrind.
738 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
739 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
740 was because its rules for how to advance for /g after matching an empty
741 string at the end of a line did not allow for this case. They now check for
744 20. pcretest is supposed to handle patterns and data of any length, by
745 extending its buffers when necessary. It was getting this wrong when the
746 buffer for a data line had to be extended.
748 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
749 CRLF as a newline sequence.
751 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
752 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
753 I have nevertheless tidied it up.
755 23. Added some casts to kill warnings from HP-UX ia64 compiler.
757 24. Added a man page for pcre-config.
760 Version 7.0 19-Dec-06
761 ---------------------
763 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
766 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
767 sys/time.h, which is documented as needed for this function. It doesn't
768 seem to matter on Linux, but it showed up on some releases of OS X.
770 3. It seems that there are systems where bytes whose values are greater than
771 127 match isprint() in the "C" locale. The "C" locale should be the
772 default when a C program starts up. In most systems, only ASCII printing
773 characters match isprint(). This difference caused the output from pcretest
774 to vary, making some of the tests fail. I have changed pcretest so that:
776 (a) When it is outputting text in the compiled version of a pattern, bytes
777 other than 32-126 are always shown as hex escapes.
779 (b) When it is outputting text that is a matched part of a subject string,
780 it does the same, unless a different locale has been set for the match
781 (using the /L modifier). In this case, it uses isprint() to decide.
783 4. Fixed a major bug that caused incorrect computation of the amount of memory
784 required for a compiled pattern when options that changed within the
785 pattern affected the logic of the preliminary scan that determines the
786 length. The relevant options are -x, and -i in UTF-8 mode. The result was
787 that the computed length was too small. The symptoms of this bug were
788 either the PCRE error "internal error: code overflow" from pcre_compile(),
789 or a glibc crash with a message such as "pcretest: free(): invalid next
790 size (fast)". Examples of patterns that provoked this bug (shown in
791 pcretest format) are:
798 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
799 is now done differently.
801 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
802 wrapper classes; (b) implement a new function in the C++ scanner that is
803 more efficient than the old way of doing things because it avoids levels of
804 recursion in the regex matching; (c) add a paragraph to the documentation
805 for the FullMatch() function.
807 6. The escape sequence \n was being treated as whatever was defined as
808 "newline". Not only was this contrary to the documentation, which states
809 that \n is character 10 (hex 0A), but it also went horribly wrong when
810 "newline" was defined as CRLF. This has been fixed.
812 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
813 was being set to -1 for the "end of line" case (supposedly a value that no
814 character can have). Though this value is never used (the check for end of
815 line is "zero bytes in current character"), it caused compiler complaints.
816 I've changed it to 0xffffffff.
818 8. In pcre_version.c, the version string was being built by a sequence of
819 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
820 string (as it is for production releases) called a macro with an empty
821 argument. The C standard says the result of this is undefined. The gcc
822 compiler treats it as an empty string (which was what was wanted) but it is
823 reported that Visual C gives an error. The source has been hacked around to
826 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
827 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
828 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
831 10. Originally, pcretest opened its input and output without "b"; then I was
832 told that "b" was needed in some environments, so it was added for release
833 5.0 to both the input and output. (It makes no difference on Unix-like
834 systems.) Later I was told that it is wrong for the input on Windows. I've
835 now abstracted the modes into two macros, to make it easier to fiddle with
836 them, and removed "b" from the input mode under Windows.
838 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
840 12. Added -help and --help to pcretest as an official way of being reminded
843 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
844 and pcrecpp.cc because they annoy compilers at high warning levels.
846 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
848 15. Fixed an occurrence of == in configure.ac that should have been = (shell
849 scripts are not C programs :-) and which was not noticed because it works
852 16. pcretest is supposed to handle any length of pattern and data line (as one
853 line or as a continued sequence of lines) by extending its input buffer if
854 necessary. This feature was broken for very long pattern lines, leading to
855 a string of junk being passed to pcre_compile() if the pattern was longer
858 17. I have done a major re-factoring of the way pcre_compile() computes the
859 amount of memory needed for a compiled pattern. Previously, there was code
860 that made a preliminary scan of the pattern in order to do this. That was
861 OK when PCRE was new, but as the facilities have expanded, it has become
862 harder and harder to keep it in step with the real compile phase, and there
863 have been a number of bugs (see for example, 4 above). I have now found a
864 cunning way of running the real compile function in a "fake" mode that
865 enables it to compute how much memory it would need, while actually only
866 ever using a few hundred bytes of working memory and without too many
867 tests of the mode. This should make future maintenance and development
868 easier. A side effect of this work is that the limit of 200 on the nesting
869 depth of parentheses has been removed (though this was never a serious
870 limitation, I suspect). However, there is a downside: pcre_compile() now
871 runs more slowly than before (30% or more, depending on the pattern). I
872 hope this isn't a big issue. There is no effect on runtime performance.
874 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
875 newline (only possible for the last line of a file) and it was a
876 pattern that set a locale (followed by /Lsomething), pcretest crashed.
878 19. Added additional timing features to pcretest. (1) The -tm option now times
879 matching only, not compiling. (2) Both -t and -tm can be followed, as a
880 separate command line item, by a number that specifies the number of
881 repeats to use when timing. The default is 50000; this gives better
882 precision, but takes uncomfortably long for very large patterns.
884 20. Extended pcre_study() to be more clever in cases where a branch of a
885 subpattern has no definite first character. For example, (a*|b*)[cd] would
886 previously give no result from pcre_study(). Now it recognizes that the
887 first character must be a, b, c, or d.
889 21. There was an incorrect error "recursive call could loop indefinitely" if
890 a subpattern (or the entire pattern) that was being tested for matching an
891 empty string contained only one non-empty item after a nested subpattern.
892 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
893 incorrectly, because the \d was being skipped in the check.
895 22. The pcretest program now has a new pattern option /B and a command line
896 option -b, which is equivalent to adding /B to every pattern. This causes
897 it to show the compiled bytecode, without the additional information that
898 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
899 is the same as /B/I).
901 23. A new optimization is now able automatically to treat some sequences such
902 as a*b as a*+b. More specifically, if something simple (such as a character
903 or a simple class like \d) has an unlimited quantifier, and is followed by
904 something that cannot possibly match the quantified thing, the quantifier
905 is automatically "possessified".
907 24. A recursive reference to a subpattern whose number was greater than 39
908 went wrong under certain circumstances in UTF-8 mode. This bug could also
909 have affected the operation of pcre_study().
911 25. Realized that a little bit of performance could be had by replacing
912 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
914 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
916 27. Possessive quantifiers such as a++ were previously implemented by turning
917 them into atomic groups such as ($>a+). Now they have their own opcodes,
918 which improves performance. This includes the automatically created ones
921 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
922 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
923 the first matched character to be a colon. This applied both to named and
926 29. The ucpinternal.h header file was missing its idempotency #ifdef.
928 30. I was sent a "project" file called libpcre.a.dev which I understand makes
929 building PCRE on Windows easier, so I have included it in the distribution.
931 31. There is now a check in pcretest against a ridiculously large number being
932 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
933 loop, the loop is abandoned.
935 32. Forward references to subpatterns in conditions such as (?(2)...) where
936 subpattern 2 is defined later cause pcre_compile() to search forwards in
937 the pattern for the relevant set of parentheses. This search went wrong
938 when there were unescaped parentheses in a character class, parentheses
939 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
941 33. "Subroutine" calls and backreferences were previously restricted to
942 referencing subpatterns earlier in the regex. This restriction has now
945 34. Added a number of extra features that are going to be in Perl 5.10. On the
946 whole, these are just syntactic alternatives for features that PCRE had
947 previously implemented using the Python syntax or my own invention. The
948 other formats are all retained for compatibility.
950 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
951 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
952 also .NET compatible.
954 (b) A recursion or subroutine call to a named group can now be defined as
955 (?&name) as well as (?P>name).
957 (c) A backreference to a named group can now be defined as \k<name> or
958 \k'name' as well as (?P=name). The new forms, as well as being in Perl
959 5.10, are also .NET compatible.
961 (d) A conditional reference to a named group can now use the syntax
962 (?(<name>) or (?('name') as well as (?(name).
964 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
965 groups (named and numbered) that are never evaluated inline, but can be
966 called as "subroutines" from elsewhere. In effect, the DEFINE condition
967 is always false. There may be only one alternative in such a group.
969 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
970 as the simple (?(R). The condition is true only if the most recent
971 recursion is that of the given number or name. It does not search out
972 through the entire recursion stack.
974 (g) The escape \gN or \g{N} has been added, where N is a positive or
975 negative number, specifying an absolute or relative reference.
977 35. Tidied to get rid of some further signed/unsigned compiler warnings and
978 some "unreachable code" warnings.
980 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
981 things, this adds five new scripts.
983 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
984 There were also incompatibilities regarding the handling of \Q..\E inside
985 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
986 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
988 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
989 matches an empty string, and forcibly breaks the loop. There were bugs in
990 this code in non-simple cases. For a pattern such as ^(a()*)* matched
991 against aaaa the result was just "a" rather than "aaaa", for example. Two
992 separate and independent bugs (that affected different cases) have been
995 39. Refactored the code to abolish the use of different opcodes for small
996 capturing bracket numbers. This is a tidy that I avoided doing when I
997 removed the limit on the number of capturing brackets for 3.5 back in 2001.
998 The new approach is not only tidier, it makes it possible to reduce the
999 memory needed to fix the previous bug (38).
1001 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1002 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1003 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1006 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1009 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1010 copy construction and assignment in the C++ wrapper.
1012 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1013 couple of bugs that could have given wrong results in the "--newline=crlf"
1016 44. Added a number of casts and did some reorganization of signed/unsigned int
1017 variables following suggestions from Dair Grant. Also renamed the variable
1018 "this" as "item" because it is a C++ keyword.
1020 45. Arranged for dftables to add
1022 #include "pcre_internal.h"
1024 to pcre_chartables.c because without it, gcc 4.x may remove the array
1025 definition from the final binary if PCRE is built into a static library and
1026 dead code stripping is activated.
1028 46. For an unanchored pattern, if a match attempt fails at the start of a
1029 newline sequence, and the newline setting is CRLF or ANY, and the next two
1030 characters are CRLF, advance by two characters instead of one.
1033 Version 6.7 04-Jul-06
1034 ---------------------
1036 1. In order to handle tests when input lines are enormously long, pcretest has
1037 been re-factored so that it automatically extends its buffers when
1038 necessary. The code is crude, but this _is_ just a test program. The
1039 default size has been increased from 32K to 50K.
1041 2. The code in pcre_study() was using the value of the re argument before
1042 testing it for NULL. (Of course, in any sensible call of the function, it
1045 3. The memmove() emulation function in pcre_internal.h, which is used on
1046 systems that lack both memmove() and bcopy() - that is, hardly ever -
1047 was missing a "static" storage class specifier.
1049 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1050 containing an extended class (one that cannot be represented by a bitmap
1051 because it contains high-valued characters or Unicode property items, e.g.
1052 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1053 pattern, but PCRE should not loop if you do not (it no longer does).
1054 [Detail: two cases were found: (a) a repeated subpattern containing an
1055 extended class; (b) a recursive reference to a subpattern that followed a
1056 previous extended class. It wasn't skipping over the extended class
1057 correctly when UTF-8 mode was not set.]
1059 5. A negated single-character class was not being recognized as fixed-length
1060 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1061 compile error "lookbehind assertion is not fixed length".
1063 6. The RunPerlTest auxiliary script was showing an unexpected difference
1064 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1065 write a Perl script that can interpret lines of an input file either as
1066 byte characters or as UTF-8, which is what "perltest" was being required to
1067 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1068 can't do is switch easily at run time between having the "use utf8;" pragma
1069 or not. In the end, I fudged it by using the RunPerlTest script to insert
1070 "use utf8;" explicitly for the UTF-8 tests.
1072 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1073 the end of the subject string, contrary to the documentation and to what
1074 Perl does. This was true of both matching functions. Now it matches only at
1075 the start of the subject and immediately after *internal* newlines.
1077 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1078 a pointer to an int instead of a pointer to an unsigned long int. This
1079 caused problems on 64-bit systems.
1081 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1082 instance of the 'standard' template library not being so standard".
1084 10. There was no check on the number of named subpatterns nor the maximum
1085 length of a subpattern name. The product of these values is used to compute
1086 the size of the memory block for a compiled pattern. By supplying a very
1087 long subpattern name and a large number of named subpatterns, the size
1088 computation could be caused to overflow. This is now prevented by limiting
1089 the length of names to 32 characters, and the number of named subpatterns
1092 11. Subpatterns that are repeated with specific counts have to be replicated in
1093 the compiled pattern. The size of memory for this was computed from the
1094 length of the subpattern and the repeat count. The latter is limited to
1095 65535, but there was no limit on the former, meaning that integer overflow
1096 could in principle occur. The compiled length of a repeated subpattern is
1097 now limited to 30,000 bytes in order to prevent this.
1099 12. Added the optional facility to have named substrings with the same name.
1101 13. Added the ability to use a named substring as a condition, using the
1102 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1103 are numbers (not recommended). Forward references are permitted.
1105 14. Added forward references in named backreferences (if you see what I mean).
1107 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1108 pattern could run off the end of the subject. For example, the pattern
1109 "(?s)(.{1,5})"8 did this with the subject "ab".
1111 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1112 PCRE_CASELESS was set when matching characters that were quantified with ?
1115 17. A character class other than a single negated character that had a minimum
1116 but no maximum quantifier - for example [ab]{6,} - was not handled
1117 correctly by pce_dfa_exec(). It would match only one character.
1119 18. A valid (though odd) pattern that looked like a POSIX character
1120 class but used an invalid character after [ (for example [[,abc,]]) caused
1121 pcre_compile() to give the error "Failed: internal error: code overflow" or
1122 in some cases to crash with a glibc free() error. This could even happen if
1123 the pattern terminated after [[ but there just happened to be a sequence of
1124 letters, a binary zero, and a closing ] in the memory that followed.
1126 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1127 over the years. Originally (before any Unicode support), just the bottom 8
1128 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1129 output from "man perlunicode" includes this:
1131 The regular expression compiler produces polymorphic opcodes. That
1132 is, the pattern adapts to the data and automatically switches to
1133 the Unicode character scheme when presented with Unicode data--or
1134 instead uses a traditional byte scheme when presented with byte
1137 Sadly, a wide octal escape does not cause a switch, and in a string with
1138 no other multibyte characters, these octal escapes are treated as before.
1139 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1140 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1143 I have not perpetrated such confusion in PCRE. Up till now, it took just
1144 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1145 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1146 translate to the appropriate multibyte character.
1148 29. Applied some refactoring to reduce the number of warnings from Microsoft
1149 and Borland compilers. This has included removing the fudge introduced
1150 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1151 a warning about an unused variable.
1153 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1154 characters since release 4.0, because Perl (from release 5.004) does not.
1155 [Or at least, is documented not to: some releases seem to be in conflict
1156 with the documentation.] However, when a pattern was studied with
1157 pcre_study() and all its branches started with \s, PCRE still included VT
1158 as a possible starting character. Of course, this did no harm; it just
1159 caused an unnecessary match attempt.
1161 22. Removed a now-redundant internal flag bit that recorded the fact that case
1162 dependency changed within the pattern. This was once needed for "required
1163 byte" processing, but is no longer used. This recovers a now-scarce options
1164 bit. Also moved the least significant internal flag bit to the most-
1165 significant bit of the word, which was not previously used (hangover from
1166 the days when it was an int rather than a uint) to free up another bit for
1169 23. Added support for CRLF line endings as well as CR and LF. As well as the
1170 default being selectable at build time, it can now be changed at runtime
1171 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1172 specify that it is scanning data with non-default line endings.
1174 24. Changed the definition of CXXLINK to make it agree with the definition of
1175 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1177 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1178 recursions. This makes a big different to stack usage for some patterns.
1180 26. If a subpattern containing a named recursion or subroutine reference such
1181 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1182 the space required for the compiled pattern went wrong and gave too small a
1183 value. Depending on the environment, this could lead to "Failed: internal
1184 error: code overflow at offset 49" or "glibc detected double free or
1187 27. Applied patches from Google (a) to support the new newline modes and (b) to
1188 advance over multibyte UTF-8 characters in GlobalReplace.
1190 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1191 difference for some implementation of PCRE in some Windows version.
1193 29. Added some extra testing facilities to pcretest:
1195 \q<number> in a data line sets the "match limit" value
1196 \Q<number> in a data line sets the "match recursion limt" value
1197 -S <number> sets the stack size, where <number> is in megabytes
1199 The -S option isn't available for Windows.
1202 Version 6.6 06-Feb-06
1203 ---------------------
1205 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1206 in pcreposix.h. I have copied the definition from pcre.h.
1208 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1209 because pcre.h is no longer a built file.
1211 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1212 not normally included in the compiled code.
1215 Version 6.5 01-Feb-06
1216 ---------------------
1218 1. When using the partial match feature with pcre_dfa_exec(), it was not
1219 anchoring the second and subsequent partial matches at the new starting
1220 point. This could lead to incorrect results. For example, with the pattern
1221 /1234/, partially matching against "123" and then "a4" gave a match.
1223 2. Changes to pcregrep:
1225 (a) All non-match returns from pcre_exec() were being treated as failures
1226 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1227 error message is output. Some extra information is given for the
1228 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1229 probably the only errors that are likely to be caused by users (by
1230 specifying a regex that has nested indefinite repeats, for instance).
1231 If there are more than 20 of these errors, pcregrep is abandoned.
1233 (b) A binary zero was treated as data while matching, but terminated the
1234 output line if it was written out. This has been fixed: binary zeroes
1235 are now no different to any other data bytes.
1237 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1238 used to set a locale for matching. The --locale=xxxx long option has
1239 been added (no short equivalent) to specify a locale explicitly on the
1240 pcregrep command, overriding the environment variables.
1242 (d) When -B was used with -n, some line numbers in the output were one less
1243 than they should have been.
1245 (e) Added the -o (--only-matching) option.
1247 (f) If -A or -C was used with -c (count only), some lines of context were
1248 accidentally printed for the final match.
1250 (g) Added the -H (--with-filename) option.
1252 (h) The combination of options -rh failed to suppress file names for files
1253 that were found from directory arguments.
1255 (i) Added the -D (--devices) and -d (--directories) options.
1257 (j) Added the -F (--fixed-strings) option.
1259 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1261 (l) Added the --colo(u)r option.
1263 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1264 is not present by default.
1266 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1267 items such as (?R) or (?1), when the recursion could match a number of
1268 alternatives. If it matched one of the alternatives, but subsequently,
1269 outside the recursion, there was a failure, the code tried to back up into
1270 the recursion. However, because of the way PCRE is implemented, this is not
1271 possible, and the result was an incorrect result from the match.
1273 In order to prevent this happening, the specification of recursion has
1274 been changed so that all such subpatterns are automatically treated as
1275 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1277 4. I had overlooked the fact that, in some locales, there are characters for
1278 which isalpha() is true but neither isupper() nor islower() are true. In
1279 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1280 and ordfeminine) are like this. This affected the treatment of \w and \W
1281 when they appeared in character classes, but not when they appeared outside
1282 a character class. The bit map for "word" characters is now created
1283 separately from the results of isalnum() instead of just taking it from the
1284 upper, lower, and digit maps. (Plus the underscore character, of course.)
1286 5. The above bug also affected the handling of POSIX character classes such as
1287 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1288 permanent tables. Instead, the bit maps for such a class were previously
1289 created as the appropriate unions of the upper, lower, and digit bitmaps.
1290 Now they are created by subtraction from the [[:word:]] class, which has
1293 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1294 It is created by subtracting the vertical space characters (\x09, \x0a,
1295 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1296 subtraction was done in the overall bitmap for a character class, meaning
1297 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1298 be recognized. This bug has been fixed.
1300 7. Patches from the folks at Google:
1302 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1303 real life, but is still worth protecting against".
1305 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1306 regular expressions".
1308 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1311 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1312 "configure" and the latter not, in order to fix a problem somebody had
1313 with compiling the Arg class on HP-UX.
1315 (e) Improve the error-handling of the C++ wrapper a little bit.
1317 (f) New tests for checking recursion limiting.
1319 8. The pcre_memmove() function, which is used only if the environment does not
1320 have a standard memmove() function (and is therefore rarely compiled),
1321 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1322 returning a result (though PCRE never actually uses the result).
1324 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1325 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1326 returned instead of calling malloc() with an overflowing number that would
1327 most likely cause subsequent chaos.
1329 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1331 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1332 with this option is matched, the nmatch and pmatch options of regexec() are
1335 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1336 provided in case anyone wants to the the POSIX interface with UTF-8
1339 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1340 C++ linking (needed for some HP-UX environments).
1342 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1343 (unused parameter) and in the pcre_printint() function (omitted "default"
1344 switch label when the default is to do nothing).
1346 15. Added some code to make it possible, when PCRE is compiled as a C++
1347 library, to replace subject pointers for pcre_exec() with a smart pointer
1348 class, thus making it possible to process discontinuous strings.
1350 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1351 much the same function. They were added by different people who were trying
1352 to make PCRE easy to compile on non-Unix systems. It has been suggested
1353 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1354 for compiling on Windows systems. I have therefore replaced it with
1355 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1356 defaults to "extern" for C or "extern C" for C++, which works fine on
1357 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1358 SCOPE with something explicit in config.h. In addition:
1360 (a) pcreposix.h still had just "extern" instead of either of these macros;
1361 I have replaced it with PCRE_DATA_SCOPE.
1363 (b) Functions such as _pcre_xclass(), which are internal to the library,
1364 but external in the C sense, all had PCRE_EXPORT in their definitions.
1365 This is apparently wrong for the Windows case, so I have removed it.
1366 (It makes no difference on Unix-like systems.)
1368 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1369 of recursive calls to match(). This is different to MATCH_LIMIT because
1370 that limits the total number of calls to match(), not all of which increase
1371 the depth of recursion. Limiting the recursion depth limits the amount of
1372 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1373 when PCRE is compiled, and changed at run time. A patch from Google adds
1374 this functionality to the C++ interface.
1376 18. Changes to the handling of Unicode character properties:
1378 (a) Updated the table to Unicode 4.1.0.
1380 (b) Recognize characters that are not in the table as "Cn" (undefined).
1382 (c) I revised the way the table is implemented to a much improved format
1383 which includes recognition of ranges. It now supports the ranges that
1384 are defined in UnicodeData.txt, and it also amalgamates other
1385 characters into ranges. This has reduced the number of entries in the
1386 table from around 16,000 to around 3,000, thus reducing its size
1387 considerably. I realized I did not need to use a tree structure after
1388 all - a binary chop search is just as efficient. Having reduced the
1389 number of entries, I extended their size from 6 bytes to 8 bytes to
1390 allow for more data.
1392 (d) Added support for Unicode script names via properties such as \p{Han}.
1394 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1395 matching that character.
1397 20. When matching a repeated Unicode property with a minimum greater than zero,
1398 (for example \pL{2,}), PCRE could look past the end of the subject if it
1399 reached it while seeking the minimum number of characters. This could
1400 happen only if some of the characters were more than one byte long, because
1401 there is a check for at least the minimum number of bytes.
1403 21. Refactored the implementation of \p and \P so as to be more general, to
1404 allow for more different types of property in future. This has changed the
1405 compiled form incompatibly. Anybody with saved compiled patterns that use
1406 \p or \P will have to recompile them.
1408 22. Added "Any" and "L&" to the supported property types.
1410 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1411 but give a compile time error if the value is greater than 0xff.
1413 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1414 accidentally not being installed or uninstalled.
1416 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1417 made were to insert the current release number. This seemed silly, because
1418 it made things harder for people building PCRE on systems that don't run
1419 "configure". I have turned pcre.h into a distributed file, no longer built
1420 by "configure", with the version identification directly included. There is
1421 no longer a pcre.h.in file.
1423 However, this change necessitated a change to the pcre-config script as
1424 well. It is built from pcre-config.in, and one of the substitutions was the
1425 release number. I have updated configure.ac so that ./configure now finds
1426 the release number by grepping pcre.h.
1428 26. Added the ability to run the tests under valgrind.
1431 Version 6.4 05-Sep-05
1432 ---------------------
1434 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1435 "--" to be printed when multiple files were scanned, even when none of the
1436 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1437 consider it to be a bug, and have restored the previous behaviour.
1439 2. A couple of code tidies to get rid of compiler warnings.
1441 3. The pcretest program used to cheat by referring to symbols in the library
1442 whose names begin with _pcre_. These are internal symbols that are not
1443 really supposed to be visible externally, and in some environments it is
1444 possible to suppress them. The cheating is now confined to including
1445 certain files from the library's source, which is a bit cleaner.
1447 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1448 file's purpose clearer.
1450 5. Reorganized pcre_ucp_findchar().
1453 Version 6.3 15-Aug-05
1454 ---------------------
1456 1. The file libpcre.pc.in did not have general read permission in the tarball.
1458 2. There were some problems when building without C++ support:
1460 (a) If C++ support was not built, "make install" and "make test" still
1463 (b) There were problems when the value of CXX was explicitly set. Some
1464 changes have been made to try to fix these, and ...
1466 (c) --disable-cpp can now be used to explicitly disable C++ support.
1468 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1469 backslash in a target when C++ was disabled. This confuses some
1470 versions of "make", apparently. Using an intermediate variable solves
1471 this. (Same for CPP_LOBJ.)
1473 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1474 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1475 necessary on certain architectures.
1477 4. Added a setting of -export-symbols-regex to the link command to remove
1478 those symbols that are exported in the C sense, but actually are local
1479 within the library, and not documented. Their names all begin with
1480 "_pcre_". This is not a perfect job, because (a) we have to except some
1481 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1482 available (and never for static libraries). I have made a note to try to
1483 find a way round (a) in the future.
1486 Version 6.2 01-Aug-05
1487 ---------------------
1489 1. There was no test for integer overflow of quantifier values. A construction
1490 such as {1111111111111111} would give undefined results. What is worse, if
1491 a minimum quantifier for a parenthesized subpattern overflowed and became
1492 negative, the calculation of the memory size went wrong. This could have
1493 led to memory overwriting.
1495 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1497 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1498 operating environments where this matters.
1500 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1501 PCRE options from within the C++ wrapper.
1503 5. Named capturing subpatterns were not being correctly counted when a pattern
1504 was compiled. This caused two problems: (a) If there were more than 100
1505 such subpatterns, the calculation of the memory needed for the whole
1506 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1507 back references of the form \12, where the number was greater than 9, were
1508 not recognized as back references, even though there were sufficient
1509 previous subpatterns.
1511 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1512 versions of gcc, e.g. 2.95.4.
1515 Version 6.1 21-Jun-05
1516 ---------------------
1518 1. There was one reference to the variable "posix" in pcretest.c that was not
1519 surrounded by "#if !defined NOPOSIX".
1521 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1522 the cross-check on the old pcre_info() function, for the benefit of the
1523 cut-down version of PCRE that is currently imported into Exim.
1525 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1526 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1527 patterns that start (?i) but I don't think that matters. The use of (?i) is
1528 just an example; this all applies to the other options as well.
1530 4. Since libtool seems to echo the compile commands it is issuing, the output
1531 from "make" can be reduced a bit by putting "@" in front of each libtool
1534 5. Patch from the folks at Google for configure.in to be a bit more thorough
1535 in checking for a suitable C++ installation before trying to compile the
1536 C++ stuff. This should fix a reported problem when a compiler was present,
1537 but no suitable headers.
1539 6. The man pages all had just "PCRE" as their title. I have changed them to
1540 be the relevant file name. I have also arranged that these names are
1541 retained in the file doc/pcre.txt, which is a concatenation in text format
1542 of all the man pages except the little individual ones for each function.
1544 7. The NON-UNIX-USE file had not been updated for the different set of source
1545 files that come with release 6. I also added a few comments about the C++
1549 Version 6.0 07-Jun-05
1550 ---------------------
1552 1. Some minor internal re-organization to help with my DFA experiments.
1554 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1555 didn't matter for the library itself when fully configured, but did matter
1556 when compiling without UCP support, or within Exim, where the ucp files are
1559 3. Refactoring of the library code to split up the various functions into
1560 different source modules. The addition of the new DFA matching code (see
1561 below) to a single monolithic source would have made it really too
1562 unwieldy, quite apart from causing all the code to be include in a
1563 statically linked application, when only some functions are used. This is
1564 relevant even without the DFA addition now that patterns can be compiled in
1565 one application and matched in another.
1567 The downside of splitting up is that there have to be some external
1568 functions and data tables that are used internally in different modules of
1569 the library but which are not part of the API. These have all had their
1570 names changed to start with "_pcre_" so that they are unlikely to clash
1571 with other external names.
1573 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1574 a different (DFA) algorithm. Although it is slower than the original
1575 function, it does have some advantages for certain types of matching
1578 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1579 including restarting after a partial match.
1581 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1582 defined when compiling for Windows was sent to me. I have put it into the
1583 code, though I have no means of testing or verifying it.
1585 7. Added the pcre_refcount() auxiliary function.
1587 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1588 match before or at the first newline in the subject string. In pcretest,
1589 the /f option on a pattern can be used to set this.
1591 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1592 would behave wrongly. This has been present in PCRE since release 4.0.
1594 10. A number of changes to the pcregrep command:
1596 (a) Refactored how -x works; insert ^(...)$ instead of setting
1597 PCRE_ANCHORED and checking the length, in preparation for adding
1598 something similar for -w.
1600 (b) Added the -w (match as a word) option.
1602 (c) Refactored the way lines are read and buffered so as to have more
1603 than one at a time available.
1605 (d) Implemented a pcregrep test script.
1607 (e) Added the -M (multiline match) option. This allows patterns to match
1608 over several lines of the subject. The buffering ensures that at least
1609 8K, or the rest of the document (whichever is the shorter) is available
1610 for matching (and similarly the previous 8K for lookbehind assertions).
1612 (f) Changed the --help output so that it now says
1616 instead of two lines, one with "regex" and the other with "regexp"
1617 because that confused at least one person since the short forms are the
1618 same. (This required a bit of code, as the output is generated
1619 automatically from a table. It wasn't just a text change.)
1621 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1622 option but starts with a hyphen. Could be a pattern or a path name
1623 starting with a hyphen, for instance.
1625 (h) "-" can be given as a file name to represent stdin.
1627 (i) When file names are being printed, "(standard input)" is used for
1628 the standard input, for compatibility with GNU grep. Previously
1631 (j) The option --label=xxx can be used to supply a name to be used for
1632 stdin when file names are being printed. There is no short form.
1634 (k) Re-factored the options decoding logic because we are going to add
1635 two more options that take data. Such options can now be given in four
1636 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1638 (l) Added the -A, -B, and -C options for requesting that lines of context
1639 around matches be printed.
1641 (m) Added the -L option to print the names of files that do not contain
1642 any matching lines, that is, the complement of -l.
1644 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1645 continue to scan other files.
1647 (o) The -s option was incorrectly implemented. For compatibility with other
1648 greps, it now suppresses the error message for a non-existent or non-
1649 accessible file (but not the return code). There is a new option called
1650 -q that suppresses the output of matching lines, which was what -s was
1653 (p) Added --include and --exclude options to specify files for inclusion
1654 and exclusion when recursing.
1656 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1657 Hopefully, it now does.
1659 12. Missing cast in pcre_study().
1661 13. Added an "uninstall" target to the makefile.
1663 14. Replaced "extern" in the function prototypes in Makefile.in with
1664 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1665 world, but is set differently for Windows.
1667 15. Added a second compiling function called pcre_compile2(). The only
1668 difference is that it has an extra argument, which is a pointer to an
1669 integer error code. When there is a compile-time failure, this is set
1670 non-zero, in addition to the error test pointer being set to point to an
1671 error message. The new argument may be NULL if no error number is required
1672 (but then you may as well call pcre_compile(), which is now just a
1673 wrapper). This facility is provided because some applications need a
1674 numeric error indication, but it has also enabled me to tidy up the way
1675 compile-time errors are handled in the POSIX wrapper.
1677 16. Added VPATH=.libs to the makefile; this should help when building with one
1678 prefix path and installing with another. (Or so I'm told by someone who
1679 knows more about this stuff than I do.)
1681 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1682 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1683 match everything, including newlines. This is not POSIX-compatible, but
1684 somebody wanted the feature. From pcretest it can be activated by using
1685 both the P and the s flags.
1687 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1689 19. libpcre.pc was being incorrectly installed as executable.
1691 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1692 it now also looks for '\r' so that it will work unmodified on Windows.
1694 21. Added Google's contributed C++ wrapper to the distribution.
1696 22. Added some untidy missing memory free() calls in pcretest, to keep
1697 Electric Fence happy when testing.
1701 Version 5.0 13-Sep-04
1702 ---------------------
1704 1. Internal change: literal characters are no longer packed up into items
1705 containing multiple characters in a single byte-string. Each character
1706 is now matched using a separate opcode. However, there may be more than one
1707 byte in the character in UTF-8 mode.
1709 2. The pcre_callout_block structure has two new fields: pattern_position and
1710 next_item_length. These contain the offset in the pattern to the next match
1711 item, and its length, respectively.
1713 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1714 insertion of callouts before each pattern item. Added the /C option to
1715 pcretest to make use of this.
1717 4. On the advice of a Windows user, the lines
1719 #if defined(_WIN32) || defined(WIN32)
1720 _setmode( _fileno( stdout ), 0x8000 );
1721 #endif /* defined(_WIN32) || defined(WIN32) */
1723 have been added to the source of pcretest. This apparently does useful
1724 magic in relation to line terminators.
1726 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1727 for the benefit of those environments where the "b" makes a difference.
1729 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1730 to know about it. I have put a hack into configure.in that adds in code
1731 to set GCC=yes if CC=icc. This seems to end up at a point in the
1732 generated configure script that is early enough to affect the setting of
1733 compiler options, which is what is needed, but I have no means of testing
1734 whether it really works. (The user who reported this had patched the
1735 generated configure script, which of course I cannot do.)
1737 LATER: After change 22 below (new libtool files), the configure script
1738 seems to know about icc (and also ecc). Therefore, I have commented out
1739 this hack in configure.in.
1741 7. Added support for pkg-config (2 patches were sent in).
1743 8. Negated POSIX character classes that used a combination of internal tables
1744 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1745 [[:^ascii]]. Typically, they would match almost any characters. The other
1746 POSIX classes were not broken in this way.
1748 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1749 to find the match, as PCRE was deluded into thinking that the match had to
1750 start at the start point or following a newline. The same bug applied to
1751 patterns with negative forward assertions or any backward assertions
1752 preceding ".*" at the start, unless the pattern required a fixed first
1753 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1755 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1756 starting at the last subject character, bytes beyond the end of the subject
1759 11. Renamed the variable "class" as "classbits" to make life easier for C++
1760 users. (Previously there was a macro definition, but it apparently wasn't
1763 12. Added the new field "tables" to the extra data so that tables can be passed
1764 in at exec time, or the internal tables can be re-selected. This allows
1765 a compiled regex to be saved and re-used at a later time by a different
1766 program that might have everything at different addresses.
1768 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1769 -R library as well as a -L library.
1771 14. The debugging options of pcretest (-d on the command line or D on a
1772 pattern) showed incorrect output for anything following an extended class
1773 that contained multibyte characters and which was followed by a quantifier.
1775 15. Added optional support for general category Unicode character properties
1776 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1777 support. It adds about 90K to the size of the library. The meanings of the
1778 inbuilt class escapes such as \d and \s have NOT been changed.
1780 16. Updated pcredemo.c to include calls to free() to release the memory for the
1783 17. The generated file chartables.c was being created in the source directory
1784 instead of in the building directory. This caused the build to fail if the
1785 source directory was different from the building directory, and was
1788 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1789 file. No doubt somebody will tell me if they don't make sense... Also added
1790 Dan Mooney's comments about building on OpenVMS.
1792 19. Added support for partial matching via the PCRE_PARTIAL option for
1793 pcre_exec() and the \P data escape in pcretest.
1795 20. Extended pcretest with 3 new pattern features:
1797 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1798 write the compiled pattern to the file whose name is "rest-of-line".
1799 This is a straight binary dump of the data, with the saved pointer to
1800 the character tables forced to be NULL. The study data, if any, is
1801 written too. After writing, pcretest reads a new pattern.
1803 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1804 compiled pattern from the given file. There must not be any
1805 occurrences of "<" in the file name (pretty unlikely); if there are,
1806 pcretest will instead treat the initial "<" as a pattern delimiter.
1807 After reading in the pattern, pcretest goes on to read data lines as
1810 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1811 and 16-bit fields in a compiled pattern, to simulate a pattern that
1812 was compiled on a host of opposite endianness.
1814 21. The pcre-exec() function can now cope with patterns that were compiled on
1815 hosts of opposite endianness, with this restriction:
1817 As for any compiled expression that is saved and used later, the tables
1818 pointer field cannot be preserved; the extra_data field in the arguments
1819 to pcre_exec() should be used to pass in a tables address if a value
1820 other than the default internal tables were used at compile time.
1822 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1823 now diagnosed as an error. Previously, most of the time, a negative number
1824 would have been treated as zero, but if in addition "ovector" was passed as
1825 NULL, a crash could occur.
1827 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1828 new versions from the libtool 1.5 distribution (the last one is a copy of
1829 a file called libtool.m4). This seems to have fixed the need to patch
1830 "configure" to support Darwin 1.3 (which I used to do). However, I still
1831 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1834 24. Changed the PCRE licence to be the more standard "BSD" licence.
1837 Version 4.5 01-Dec-03
1838 ---------------------
1840 1. There has been some re-arrangement of the code for the match() function so
1841 that it can be compiled in a version that does not call itself recursively.
1842 Instead, it keeps those local variables that need separate instances for
1843 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1844 needs to "recurse". Keeping track of where control must go is done by means
1845 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1846 hide most of the details from the main code, and operates only if
1847 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1848 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1851 To make it easier for callers to provide specially tailored get/free
1852 functions for this usage, two new functions, pcre_stack_malloc, and
1853 pcre_stack_free, are used. They are always called in strict stacking order,
1854 and the size of block requested is always the same.
1856 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1857 PCRE has been compiled to use the stack or the heap for recursion. The
1858 -C option of pcretest uses this to show which version is compiled.
1860 A new data escape \S, is added to pcretest; it causes the amounts of store
1861 obtained and freed by both kinds of malloc/free at match time to be added
1864 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1865 what's available on my current Linux desktop machine.
1867 3. When matching a UTF-8 string, the test for a valid string at the start has
1868 been extended. If start_offset is not zero, PCRE now checks that it points
1869 to a byte that is the start of a UTF-8 character. If not, it returns
1870 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1871 this is necessary because there may be backward assertions in the pattern.
1872 When matching the same subject several times, it may save resources to use
1873 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1875 4. The code for checking the validity of UTF-8 strings has been tightened so
1876 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1877 containing "overlong sequences".
1879 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1880 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1881 should have been "&", but it just so happened that all the cases this let
1882 through by mistake were picked up later in the function.
1884 6. I had used a variable called "isblank" - this is a C99 function, causing
1885 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1887 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1888 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1889 so the version is shown; (c) stop "make test" echoing "./RunTest".
1891 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1893 9. The prototype for memmove() for systems that don't have it was using
1894 size_t, but the inclusion of the header that defines size_t was later. I've
1895 moved the #includes for the C headers earlier to avoid this.
1897 10. Added some adjustments to the code to make it easier to compiler on certain
1900 (a) Some "const" qualifiers were missing.
1901 (b) Added the macro EXPORT before all exported functions; by default this
1902 is defined to be empty.
1903 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1904 that it reads its output file name as an argument instead of writing
1905 to the standard output and assuming this can be redirected.
1907 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1908 class containing characters with values greater than 255, PCRE compilation
1911 12. A recursive reference to a subpattern that was within another subpattern
1912 that had a minimum quantifier of zero caused PCRE to crash. For example,
1913 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1914 recursion. If the recursively-called subpattern itself had a zero repeat,
1917 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1918 buffer into which it was copied (for escape processing) was still set at
1919 1024, so long lines caused crashes.
1921 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1922 "internal error: code overflow...". This applied to any character class
1923 that was followed by a possessive quantifier.
1925 15. Modified the Makefile to add libpcre.la as a prerequisite for
1926 libpcreposix.la because I was told this is needed for a parallel build to
1929 16. If a pattern that contained .* following optional items at the start was
1930 studied, the wrong optimizing data was generated, leading to matching
1931 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1932 matching string must start with a or b or c. The correct conclusion for
1933 this pattern is that a match can start with any character.
1936 Version 4.4 13-Aug-03
1937 ---------------------
1939 1. In UTF-8 mode, a character class containing characters with values between
1940 127 and 255 was not handled correctly if the compiled pattern was studied.
1941 In fixing this, I have also improved the studying algorithm for such
1944 2. Three internal functions had redundant arguments passed to them. Removal
1945 might give a very teeny performance improvement.
1947 3. Documentation bug: the value of the capture_top field in a callout is *one
1948 more than* the number of the hightest numbered captured substring.
1950 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1951 in incorrectly linking with a previously installed version. They now link
1952 explicitly with libpcre.la.
1954 5. configure.in no longer needs to recognize Cygwin specially.
1956 6. A problem in pcre.in for Windows platforms is fixed.
1958 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1959 pcretest, it used to include the size of the study block as part of its
1960 output. Unfortunately, the structure contains a field that has a different
1961 size on different hardware architectures. This meant that the tests that
1962 showed this size failed. As the block is currently always of a fixed size,
1963 this information isn't actually particularly useful in pcretest output, so
1964 I have just removed it.
1966 8. Three pre-processor statements accidentally did not start in column 1.
1967 Sadly, there are *still* compilers around that complain, even though
1968 standard C has not required this for well over a decade. Sigh.
1970 9. In pcretest, the code for checking callouts passed small integers in the
1971 callout_data field, which is a void * field. However, some picky compilers
1972 complained about the casts involved for this on 64-bit systems. Now
1973 pcretest passes the address of the small integer instead, which should get
1974 rid of the warnings.
1976 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1977 both compile and run time, and gives an error if an invalid UTF-8 sequence
1978 is found. There is a option for disabling this check in cases where the
1979 string is known to be correct and/or the maximum performance is wanted.
1981 11. In response to a bug report, I changed one line in Makefile.in from
1983 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1985 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1987 to look similar to other lines, but I have no way of telling whether this
1988 is the right thing to do, as I do not use Windows. No doubt I'll get told
1992 Version 4.3 21-May-03
1993 ---------------------
1995 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1998 2. Some refactoring to improve the quality of the code:
2000 (i) The utf8_table... variables are now declared "const".
2002 (ii) The code for \cx, which used the "case flipping" table to upper case
2003 lower case letters, now just substracts 32. This is ASCII-specific,
2004 but the whole concept of \cx is ASCII-specific, so it seems
2007 (iii) PCRE was using its character types table to recognize decimal and
2008 hexadecimal digits in the pattern. This is silly, because it handles
2009 only 0-9, a-f, and A-F, but the character types table is locale-
2010 specific, which means strange things might happen. A private
2011 table is now used for this - though it costs 256 bytes, a table is
2012 much faster than multiple explicit tests. Of course, the standard
2013 character types table is still used for matching digits in subject
2016 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2017 ending in _t are). So I've renamed it as ESC_tee.
2019 3. The first argument for regexec() in the POSIX wrapper should have been
2022 4. Changed pcretest to use malloc() for its buffers so that they can be
2023 Electric Fenced for debugging.
2025 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2026 to read one or more bytes before the start of the subject string. Often this
2027 had no effect on PCRE's behaviour, but in some circumstances it could
2028 provoke a segmentation fault.
2030 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2031 to try to read one or more bytes before the start of the subject string.
2033 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2034 UTF-8 support could misbehave in various ways if the subject string
2035 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2036 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2037 back over UTF-8 characters.)
2040 Version 4.2 14-Apr-03
2041 ---------------------
2043 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2045 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2046 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2047 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2048 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2049 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2051 Note: automatic setting of the BUILD variables is not yet working
2052 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2053 compile-time but not at link-time
2054 [LINK]: use for linking executables only
2055 make different versions for Windows and non-Windows
2056 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2058 [LINK_FOR_BUILD]: new variable
2059 [OBJEXT]: use throughout
2060 [EXEEXT]: use throughout
2061 <winshared>: new target
2062 <wininstall>: new target
2063 <dftables.o>: use native compiler
2064 <dftables>: use native linker
2065 <install>: handle Windows platform correctly
2068 copy DLL to top builddir before testing
2070 As part of these changes, -no-undefined was removed again. This was reported
2071 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2074 3. Some tidies to get rid of compiler warnings:
2076 . In the match_data structure, match_limit was an unsigned long int, whereas
2077 match_call_count was an int. I've made them both unsigned long ints.
2079 . In pcretest the fact that a const uschar * doesn't automatically cast to
2080 a void * provoked a warning.
2082 . Turning on some more compiler warnings threw up some "shadow" variables
2083 and a few more missing casts.
2085 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2086 option, a class that contained a single character with a value between 128
2087 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2089 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2090 option, a class that contained several characters, but with at least one
2091 whose value was between 128 and 255 caused PCRE to crash.
2094 Version 4.1 12-Mar-03
2095 ---------------------
2097 1. Compiling with gcc -pedantic found a couple of places where casts were
2098 needed, and a string in dftables.c that was longer than standard compilers are
2099 required to support.
2101 2. Compiling with Sun's compiler found a few more places where the code could
2102 be tidied up in order to avoid warnings.
2104 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2105 first of these names is deprecated in the latest Autoconf in favour of the name
2106 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2107 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2108 analogy I have changed it to CFLAGS_FOR_BUILD.
2110 4. Added -no-undefined to the linking command in the Makefile, because this is
2111 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2112 linking step for the pcreposix library.
2114 5. PCRE was failing to diagnose the case of two named groups with the same
2117 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2118 literal character that is needed in the subject for a match, and scans along to
2119 ensure that it is present before embarking on the full matching process. This
2120 saves time in cases of nested unlimited repeats that are never going to match.
2121 Problem: the scan can take a lot of time if the subject is very long (e.g.
2122 megabytes), thus penalizing straightforward matches. It is now done only if the
2123 amount of subject to be scanned is less than 1000 bytes.
2125 7. A lesser problem with the same optimization is that it was recording the
2126 first character of an anchored pattern as "needed", thus provoking a search
2127 right along the subject, even when the first match of the pattern was going to
2128 fail. The "needed" character is now not set for anchored patterns, unless it
2129 follows something in the pattern that is of non-fixed length. Thus, it still
2130 fulfils its original purpose of finding quick non-matches in cases of nested
2131 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2134 Version 4.0 17-Feb-03
2135 ---------------------
2137 1. If a comment in an extended regex that started immediately after a meta-item
2138 extended to the end of string, PCRE compiled incorrect data. This could lead to
2139 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2141 2. Moved to autoconf 2.53 and libtool 1.4.2.
2143 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2144 the special perltest8 script is no longer needed - all the tests can be run
2145 from a single perltest script.
2147 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2148 by \s. It has now been removed in PCRE. This means it isn't recognized as
2149 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2150 class [:space:] *does* include VT, thereby creating a mess.
2152 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2155 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2156 its new features into the main test script, reducing the number of scripts.
2158 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2159 were backward compatible, and made the (?i) apply to the whole pattern, as if
2160 /i were given. Now it behaves more logically, and applies the option setting
2161 only to what follows. PCRE has been changed to follow suit. However, if it
2162 finds options settings right at the start of the pattern, it extracts them into
2163 the global options, as before. Thus, they show up in the info data.
2165 8. Added support for the \Q...\E escape sequence. Characters in between are
2166 treated as literals. This is slightly different from Perl in that $ and @ are
2167 also handled as literals inside the quotes. In Perl, they will cause variable
2168 interpolation. Note the following examples:
2170 Pattern PCRE matches Perl matches
2172 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2173 \Qabc\$xyz\E abc\$xyz abc\$xyz
2174 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2176 For compatibility with Perl, \Q...\E sequences are recognized inside character
2177 classes as well as outside them.
2179 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2180 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2181 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2182 signed/unsigned warnings.
2184 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2185 option for pcretest, so I've replaced it by a simple function that does just
2188 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2191 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2192 Java package. This provides some syntactic sugar for simple cases of what my
2193 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2194 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2195 item, you can use this simplified notation. Note that only makes sense with
2196 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2197 greediness, whatever the setting of the PCRE_UNGREEDY option.
2199 13. A change of greediness default within a pattern was not taking effect at
2200 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2201 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2202 was abstracted outside.
2204 14. PCRE now supports the \G assertion. It is true when the current matching
2205 position is at the start point of the match. This differs from \A when the
2206 starting offset is non-zero. Used with the /g option of pcretest (or similar
2207 code), it works in the same way as it does for Perl's /g option. If all
2208 alternatives of a regex begin with \G, the expression is anchored to the start
2209 match position, and the "anchored" flag is set in the compiled expression.
2211 15. Some bugs concerning the handling of certain option changes within patterns
2212 have been fixed. These applied to options other than (?ims). For example,
2213 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2214 been the other way round. Some of this was related to change 7 above.
2216 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2217 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2218 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2219 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2221 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2222 mode. Unlike ".", it always matches newline, whatever the setting of
2223 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2224 assertions. Perl allows it, but it doesn't (in general) work because it can't
2225 calculate the length of the lookbehind. At least, that's the case for Perl
2226 5.8.0 - I've been told they are going to document that it doesn't work in
2229 18. Added an error diagnosis for escapes that PCRE does not support: these are
2230 \L, \l, \N, \P, \p, \U, \u, and \X.
2232 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2233 reading past the end of the pattern in cases such as /[abcd/.
2235 20. PCRE was getting more memory than necessary for patterns with classes that
2236 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2238 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2239 compiling PCRE for use with Virtual Pascal.
2241 22. Small fix to the Makefile to make it work properly if the build is done
2242 outside the source tree.
2244 23. Added a new extension: a condition to go with recursion. If a conditional
2245 subpattern starts with (?(R) the "true" branch is used if recursion has
2246 happened, whereas the "false" branch is used only at the top level.
2248 24. When there was a very long string of literal characters (over 255 bytes
2249 without UTF support, over 250 bytes with UTF support), the computation of how
2250 much memory was required could be incorrect, leading to segfaults or other
2253 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2254 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2255 there was a subsequent back reference to those brackets. This meant that, for
2256 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2257 possible to check for precisely this case. All we can do is abandon the
2258 optimization if .* occurs inside capturing brackets when there are any back
2259 references whatsoever. (See below for a better fix that came later.)
2261 26. The handling of the optimization for finding the first character of a
2262 non-anchored pattern, and for finding a character that is required later in the
2263 match were failing in some cases. This didn't break the matching; it just
2264 failed to optimize when it could. The way this is done has been re-implemented.
2266 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2268 28. Added a new feature that provides some of the functionality that Perl
2269 provides with (?{...}). The facility is termed a "callout". The way it is done
2270 in PCRE is for the caller to provide an optional function, by setting
2271 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2272 global variable. By default it is unset, which disables all calling out. To get
2273 the function called, the regex must include (?C) at appropriate points. This
2274 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2275 This provides a means of identifying different callout points. When PCRE
2276 reaches such a point in the regex, if pcre_callout has been set, the external
2277 function is called. It is provided with data in a structure called
2278 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2279 matching continues; if it returns a non-zero value, the match at the current
2280 point fails. However, backtracking will occur if possible. [This was changed
2281 later and other features added - see item 49 below.]
2283 29. pcretest is upgraded to test the callout functionality. It provides a
2284 callout function that displays information. By default, it shows the start of
2285 the match and the current position in the text. There are some new data escapes
2286 to vary what happens:
2288 \C+ in addition, show current contents of captured substrings
2289 \C- do not supply a callout function
2290 \C!n return 1 when callout number n is reached
2291 \C!n!m return 1 when callout number n is reached for the mth time
2293 30. If pcregrep was called with the -l option and just a single file name, it
2294 output "<stdin>" if a match was found, instead of the file name.
2296 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2297 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2298 pcre_exec(). This saves a malloc/free per call. The default value of
2299 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2302 32. The default maximum size of a compiled pattern is 64K. There have been a
2303 few cases of people hitting this limit. The code now uses macros to handle the
2304 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2305 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2306 configuring. Tests 2 and 5 work only with 2-byte links because they output
2307 debugging information about compiled patterns.
2309 33. Internal code re-arrangements:
2311 (a) Moved the debugging function for printing out a compiled regex into
2312 its own source file (printint.c) and used #include to pull it into
2313 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2316 (b) Defined the list of op-code names for debugging as a macro in
2317 internal.h so that it is next to the definition of the opcodes.
2319 (c) Defined a table of op-code lengths for simpler skipping along compiled
2320 code. This is again a macro in internal.h so that it is next to the
2321 definition of the opcodes.
2323 34. Added support for recursive calls to individual subpatterns, along the
2324 lines of Robin Houston's patch (but implemented somewhat differently).
2326 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2327 allow it to read and process whole directories in Win32. This code was
2328 contributed by Lionel Fourquaux; it has not been tested by me.
2330 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2331 used to name a group. Names consist of alphanumerics and underscores, and must
2332 be unique. Back references use the syntax (?P=name) and recursive calls use
2333 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2334 numbers. The function pcre_fullinfo() can be used after compilation to extract
2335 a name/number map. There are three relevant calls:
2337 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2338 PCRE_INFO_NAMECOUNT yields the number of entries
2339 PCRE_INFO_NAMETABLE yields a pointer to the map.
2341 The map is a vector of fixed-size entries. The size of each entry depends on
2342 the length of the longest name used. The first two bytes of each entry are the
2343 group number, most significant byte first. There follows the corresponding
2344 name, zero terminated. The names are in alphabetical order.
2346 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2347 case instead of 255. Making it the same both with and without UTF-8 support
2348 means that the same test output works with both.
2350 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2351 calling malloc() with a zero argument.
2353 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2354 optimization. I've improved things by keeping a bitmap of backreferences with
2355 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2356 fact referenced, the optimization can be applied. It is unlikely that a
2357 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2358 the match to follow \n) will appear inside brackets with a number greater than
2359 31, but if it does, any back reference > 31 suppresses the optimization.
2361 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2362 of disabling numbered capturing parentheses. Any opening parenthesis that is
2363 not followed by ? behaves as if it were followed by ?: but named parentheses
2364 can still be used for capturing (and they will acquire numbers in the usual
2367 41. Redesigned the return codes from the match() function into yes/no/error so
2368 that errors can be passed back from deep inside the nested calls. A malloc
2369 failure while inside a recursive subpattern call now causes the
2370 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2372 42. It is now possible to set a limit on the number of times the match()
2373 function is called in a call to pcre_exec(). This facility makes it possible to
2374 limit the amount of recursion and backtracking, though not in a directly
2375 obvious way, because the match() function is used in a number of different
2376 circumstances. The count starts from zero for each position in the subject
2377 string (for non-anchored patterns). The default limit is, for compatibility, a
2378 large number, namely 10 000 000. You can change this in two ways:
2380 (a) When configuring PCRE before making, you can use --with-match-limit=n
2381 to set a default value for the compiled library.
2383 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2384 a different value is set. See 45 below.
2386 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2388 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2389 of things that can be changed at compile time. The first argument specifies
2390 what is wanted and the second points to where the information is to be placed.
2391 The current list of available information is:
2395 The output is an integer that is set to one if UTF-8 support is available;
2396 otherwise it is set to zero.
2400 The output is an integer that it set to the value of the code that is used for
2401 newline. It is either LF (10) or CR (13).
2403 PCRE_CONFIG_LINK_SIZE
2405 The output is an integer that contains the number of bytes used for internal
2406 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2408 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2410 The output is an integer that contains the threshold above which the POSIX
2411 interface uses malloc() for output vectors. See item 31 above.
2413 PCRE_CONFIG_MATCH_LIMIT
2415 The output is an unsigned integer that contains the default limit of the number
2416 of match() calls in a pcre_exec() execution. See 42 above.
2418 44. pcretest has been upgraded by the addition of the -C option. This causes it
2419 to extract all the available output from the new pcre_config() function, and to
2420 output it. The program then exits immediately.
2422 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2423 order to support additional features. One way would have been to define
2424 pcre_exec2() (for example) with extra arguments, but this would not have been
2425 extensible, and would also have required all calls to the original function to
2426 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2427 is used for passing in "extra" data from pcre_study().
2429 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2430 contains the following fields:
2432 flags a bitmap indicating which of the following fields are set
2433 study_data opaque data from pcre_study()
2434 match_limit a way of specifying a limit on match() calls for a specific
2436 callout_data data for callouts (see 49 below)
2438 The flag bits are also defined in pcre.h, and are
2440 PCRE_EXTRA_STUDY_DATA
2441 PCRE_EXTRA_MATCH_LIMIT
2442 PCRE_EXTRA_CALLOUT_DATA
2444 The pcre_study() function now returns one of these new pcre_extra blocks, with
2445 the actual study data pointed to by the study_data field, and the
2446 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2447 before. That is, this change is entirely upwards-compatible and requires no
2448 change to existing code.
2450 If you want to pass in additional data to pcre_exec(), you can either place it
2451 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2454 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2455 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2456 times with different match limits, until it finds the minimum value needed for
2457 pcre_exec() to complete. The value is then output. This can be instructive; for
2458 most simple matches the number is quite small, but for pathological cases it
2459 gets very large very quickly.
2461 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2462 returns the size of the data block pointed to by the study_data field in a
2463 pcre_extra block, that is, the value that was passed as the argument to
2464 pcre_malloc() when PCRE was getting memory in which to place the information
2465 created by pcre_study(). The fourth argument should point to a size_t variable.
2466 pcretest has been extended so that this information is shown after a successful
2467 pcre_study() call when information about the compiled regex is being displayed.
2469 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2470 because what follows is always an absolute path. (Later: it turns out that this
2471 is more than cosmetic for MinGW, because it doesn't like empty path
2474 49. Some changes have been made to the callout feature (see 28 above):
2476 (i) A callout function now has three choices for what it returns:
2478 0 => success, carry on matching
2479 > 0 => failure at this point, but backtrack if possible
2480 < 0 => serious error, return this value from pcre_exec()
2482 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2483 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2484 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2485 use by callout functions. It will never be used by PCRE itself.
2487 (ii) The pcre_extra structure (see 45 above) has a void * field called
2488 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2489 pcre_callout_block structure has a field of the same name. The contents of
2490 the field passed in the pcre_extra structure are passed to the callout
2491 function in the corresponding field in the callout block. This makes it
2492 easier to use the same callout-containing regex from multiple threads. For
2493 testing, the pcretest program has a new data escape
2495 \C*n pass the number n (may be negative) as callout_data
2497 If the callout function in pcretest receives a non-zero value as
2498 callout_data, it returns that value.
2500 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2501 there were some redundant $(CFLAGS) in commands that are now specified as
2502 $(LINK), which already includes $(CFLAGS).
2504 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2505 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2506 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2507 one-byte characters throughout. Note that case-insensitive matching applies
2508 only to characters whose values are less than 256. PCRE doesn't support the
2509 notion of cases for higher-valued characters.
2511 (i) A character class whose characters are all within 0-255 is handled as
2512 a bit map, and the map is inverted for negative classes. Previously, a
2513 character > 255 always failed to match such a class; however it should
2514 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2516 (ii) A negated character class with a single character < 255 is coded as
2517 "not this character" (OP_NOT). This wasn't working properly when the test
2518 character was multibyte, either singly or repeated.
2520 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2521 mode, for example: \x{100}{2,3}.
2523 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2524 singly or repeated) now correctly test multibyte characters. However,
2525 PCRE doesn't recognize any characters with values greater than 255 as
2526 digits, spaces, or word characters. Such characters always match \D, \S,
2527 and \W, and never match \d, \s, or \w.
2529 (v) Classes may now contain characters and character ranges with values
2530 greater than 255. For example: [ab\x{100}-\x{400}].
2532 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2535 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2536 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2537 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2540 53. The single man page has become too large. I have therefore split it up into
2541 a number of separate man pages. These also give rise to individual HTML pages;
2542 these are now put in a separate directory, and there is an index.html page that
2543 lists them all. Some hyperlinking between the pages has been installed.
2545 54. Added convenience functions for handling named capturing parentheses.
2547 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2548 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2549 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2552 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2553 calling configure. These values are used when compiling the dftables.c program
2554 which is run to generate the source of the default character tables. They
2555 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2556 you will need to set these values.
2558 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2561 Version 3.9 02-Jan-02
2562 ---------------------
2564 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2566 2. If --disable-static was given, the building process failed when trying to
2567 build pcretest and pcregrep. (For some reason it was using libtool to compile
2568 them, which is not right, as they aren't part of the library.)
2571 Version 3.8 18-Dec-01
2572 ---------------------
2574 1. The experimental UTF-8 code was completely screwed up. It was packing the
2575 bytes in the wrong order. How dumb can you get?
2578 Version 3.7 29-Oct-01
2579 ---------------------
2581 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2582 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2583 this didn't happen under Solaris 8, where I normally test things.
2585 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2586 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2587 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2588 this fix an hour or so after the initial 3.7 release.)
2591 Version 3.6 23-Oct-01
2592 ---------------------
2594 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2595 offsets passed as NULL with zero offset count.
2597 2. The config.guess and config.sub files had not been updated when I moved to
2598 the latest autoconf.
2601 Version 3.5 15-Aug-01
2602 ---------------------
2604 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2607 2. By using declared but undefined structures, we can avoid using "void"
2608 definitions in pcre.h while keeping the internal definitions of the structures
2611 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2612 user point of view, this means that both static and shared libraries are built
2613 by default, but this can be individually controlled. More of the work of
2614 handling this static/shared cases is now inside libtool instead of PCRE's make
2617 4. The pcretest utility is now installed along with pcregrep because it is
2618 useful for users (to test regexs) and by doing this, it automatically gets
2619 relinked by libtool. The documentation has been turned into a man page, so
2620 there are now .1, .txt, and .html versions in /doc.
2622 5. Upgrades to pcregrep:
2623 (i) Added long-form option names like gnu grep.
2624 (ii) Added --help to list all options with an explanatory phrase.
2625 (iii) Added -r, --recursive to recurse into sub-directories.
2626 (iv) Added -f, --file to read patterns from a file.
2628 6. pcre_exec() was referring to its "code" argument before testing that
2629 argument for NULL (and giving an error if it was NULL).
2631 7. Upgraded Makefile.in to allow for compiling in a different directory from
2632 the source directory.
2634 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2635 options bits, the pointer it was passed was to an int instead of to an unsigned
2636 long int. This mattered only on 64-bit systems.
2638 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2639 generated) instead of pcre.in, which it its source. Also made the same change
2640 in several of the .c files.
2642 10. A new release of gcc defines printf() as a macro, which broke pcretest
2643 because it had an ifdef in the middle of a string argument for printf(). Fixed
2644 by using separate calls to printf().
2646 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2647 script, to force use of CR or LF instead of \n in the source. On non-Unix
2648 systems, the value can be set in config.h.
2650 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2651 absolute limit. Changed the text of the error message to make this clear, and
2652 likewise updated the man page.
2654 13. The limit of 99 on the number of capturing subpatterns has been removed.
2655 The new limit is 65535, which I hope will not be a "real" limit.
2658 Version 3.4 22-Aug-00
2659 ---------------------
2661 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2663 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2666 Version 3.3 01-Aug-00
2667 ---------------------
2669 1. If an octal character was given, but the value was greater than \377, it
2670 was not getting masked to the least significant bits, as documented. This could
2671 lead to crashes in some systems.
2673 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2674 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2676 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2677 These just pass their arguments on to (pcre_free)(), but they are provided
2678 because some uses of PCRE bind it to non-C systems that can call its functions,
2679 but cannot call free() or pcre_free() directly.
2681 4. Add "make test" as a synonym for "make check". Corrected some comments in
2684 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2687 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2688 command called pgrep for grepping around the active processes.
2690 7. Added the beginnings of support for UTF-8 character strings.
2692 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2693 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2694 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2695 out for the ar command.)
2698 Version 3.2 12-May-00
2699 ---------------------
2701 This is purely a bug fixing release.
2703 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2704 of ZA. This was just one example of several cases that could provoke this bug,
2705 which was introduced by change 9 of version 2.00. The code for breaking
2706 infinite loops after an iteration that matches an empty string was't working
2709 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2710 when matched against abbab (for example). After matching an empty string, it
2711 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2712 caused it to match further down the string than it should.
2714 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2715 was there because it doesn't seem to be needed, and it causes trouble on some
2716 systems, as it is not a Standard C header. It has been removed.
2718 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2719 were reported on the Macintosh. The changes were from
2721 while ((c = *(++ptr)) != 0 && c != '\n');
2723 while ((c = *(++ptr)) != 0 && c != '\n') ;
2725 Totally extraordinary, but if that's what it takes...
2727 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2728 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2729 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2730 assumes the way PCRE uses memmove() (always moving upwards).
2732 6. PCRE is being used in one environment where strchr() is not available. There
2733 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2737 Version 3.1 09-Feb-00
2738 ---------------------
2740 The only change in this release is the fixing of some bugs in Makefile.in for
2741 the "install" target:
2743 (1) It was failing to install pcreposix.h.
2745 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2748 Version 3.0 01-Feb-00
2749 ---------------------
2751 1. Add support for the /+ modifier to perltest (to output $` like it does in
2754 2. Add support for the /g modifier to perltest.
2756 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2757 matches null strings.
2759 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2760 pattern. Perl treats empty patterns specially - it reuses the most recent
2761 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2764 5. The POSIX interface was broken in that it was just handing over the POSIX
2765 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2766 required a bigger vector, with some working space on the end. This means that
2767 the POSIX wrapper now has to get and free some memory, and copy the results.
2769 6. Added some simple autoconf support, placing the test data and the
2770 documentation in separate directories, re-organizing some of the
2771 information files, and making it build pcre-config (a GNU standard). Also added
2772 libtool support for building PCRE as a shared library, which is now the
2775 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2776 09 are not valid octal constants. Single digits will be used for minor values
2779 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2780 existing programs that set these in the POSIX interface can use PCRE without
2783 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2784 return all that pcre_info() returns, plus additional data. The pcre_info()
2785 function is retained for compatibility, but is considered to be obsolete.
2787 10. Added experimental recursion feature (?R) to handle one common case that
2788 Perl 5.6 will be able to do with (?p{...}).
2790 11. Added support for POSIX character classes like [:alpha:], which Perl is
2794 Version 2.08 31-Aug-99
2795 ----------------------
2797 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2798 trying to match at the startoffset position, but instead was moving forward to
2799 the next newline as if a previous match had failed.
2801 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2802 and could get into a loop if a null string was matched other than at the start
2805 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2806 be distinguished at compile time, and for completeness also added PCRE_DATE.
2808 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2809 in GnuWin32 environments.
2812 Version 2.07 29-Jul-99
2813 ----------------------
2815 1. The documentation is now supplied in plain text form and HTML as well as in
2816 the form of man page sources.
2818 2. C++ compilers don't like assigning (void *) values to other pointer types.
2819 In particular this affects malloc(). Although there is no problem in Standard
2820 C, I've put in casts to keep C++ compilers happy.
2822 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2823 should be (const char *).
2825 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2826 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2827 However, I haven't made this a standard facility. The documentation doesn't
2828 mention it, and the Makefile doesn't support it.
2830 5. The Makefile now contains an "install" target, with editable destinations at
2831 the top of the file. The pcretest program is not installed.
2833 6. pgrep -V now gives the PCRE version number and date.
2835 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2836 causing the entire string to be ignored, instead of just the last character.
2838 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2839 non-matching string, it can take a very, very long time, even for strings of
2840 quite modest length, because of the nested recursion. PCRE now does better in
2841 some of these cases. It does this by remembering the last required literal
2842 character in the pattern, and pre-searching the subject to ensure it is present
2843 before running the real match. In other words, it applies a heuristic to detect
2844 some types of certain failure quickly, and in the above example, if presented
2845 with a string that has no trailing " it gives "no match" very quickly.
2847 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2848 other alternatives are tried instead.
2851 Version 2.06 09-Jun-99
2852 ----------------------
2854 1. Change pcretest's output for amount of store used to show just the code
2855 space, because the remainder (the data block) varies in size between 32-bit and
2858 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2859 start matching at. This allows lookbehinds to work when searching for multiple
2860 occurrences in a string.
2862 3. Added additional options to pcretest for testing multiple occurrences:
2864 /+ outputs the rest of the string that follows a match
2865 /g loops for multiple occurrences, using the new startoffset argument
2866 /G loops for multiple occurrences by passing an incremented pointer
2868 4. PCRE wasn't doing the "first character" optimization for patterns starting
2869 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2870 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2871 the letter 'x'. On long subject strings, this gives a significant speed-up.
2874 Version 2.05 21-Apr-99
2875 ----------------------
2877 1. Changed the type of magic_number from int to long int so that it works
2878 properly on 16-bit systems.
2880 2. Fixed a bug which caused patterns starting with .* not to work correctly
2881 when the subject string contained newline characters. PCRE was assuming
2882 anchoring for such patterns in all cases, which is not correct because .* will
2883 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2884 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2885 must be retried after every newline in the subject.
2888 Version 2.04 18-Feb-99
2889 ----------------------
2891 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2892 computation of the store needed to hold the pattern was incorrect (too large).
2893 If such patterns were nested a few deep, this could multiply and become a real
2896 2. Added /M option to pcretest to show the memory requirement of a specific
2897 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2899 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2900 compiled in such a way that the backtracking after subsequent failure was
2901 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2902 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2905 Version 2.03 02-Feb-99
2906 ----------------------
2908 1. Fixed typo and small mistake in man page.
2910 2. Added 4th condition (GPL supersedes if conflict) and created separate
2911 LICENCE file containing the conditions.
2913 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2914 Perl, that is the internal \ allows the delimiter to be included in the
2915 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2916 the final delimiter, add \ to the end of the pattern (to test the error).
2918 4. Added the convenience functions for extracting substrings after a successful
2919 match. Updated pcretest to make it able to test these functions.
2922 Version 2.02 14-Jan-99
2923 ----------------------
2925 1. Initialized the working variables associated with each extraction so that
2926 their saving and restoring doesn't refer to uninitialized store.
2928 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2929 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2932 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2933 calls, and wasn't printing the correct value for compiling calls. Increased the
2934 default value of LOOPREPEAT, and the number of significant figures in the
2937 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2939 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2940 a building problem on Windows NT with a FAT file system.
2943 Version 2.01 21-Oct-98
2944 ----------------------
2946 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2947 to character tables built by pcre_maketables() in the current locale. If NULL
2948 is passed, the default tables are used.
2951 Version 2.00 24-Sep-98
2952 ----------------------
2954 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2957 2. Allow quantification of (?>) groups, and make it work correctly.
2959 3. The first character computation wasn't working for (?>) groups.
2961 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2962 end of the subject) and add 5.005's \z, which really does match only at the
2963 very end of the subject.
2965 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2967 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2968 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2969 localized options. All options to pcre_study() were also removed.
2971 7. Add other new features from 5.005:
2973 $(?<= positive lookbehind
2974 $(?<! negative lookbehind
2975 (?imsx-imsx) added the unsetting capability
2976 such a setting is global if at outer level; local otherwise
2977 (?imsx-imsx:) non-capturing groups with option setting
2978 (?(cond)re|re) conditional pattern matching
2980 A backreference to itself in a repeated group matches the previous
2983 8. General tidying up of studying (both automatic and via "study")
2984 consequential on the addition of new assertions.
2986 9. As in 5.005, unlimited repeated groups that could match an empty substring
2987 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2988 runtime if any iteration does actually match an empty substring.
2990 10. Include the RunTest script in the distribution.
2992 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2993 discrepancies, some of which were old and were also with respect to 5.004. They
2994 have now been fixed.
2997 Version 1.09 28-Apr-98
2998 ----------------------
3000 1. A negated single character class followed by a quantifier with a minimum
3001 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3002 program crashes, or just wrong answers. This did not apply to negated classes
3003 containing more than one character, or to minima other than one.
3006 Version 1.08 27-Mar-98
3007 ----------------------
3009 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3011 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3012 latter must appear before anything that relies on it in the pattern.
3015 Version 1.07 16-Feb-98
3016 ----------------------
3018 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3019 repeat of a potentially empty string).
3022 Version 1.06 23-Jan-98
3023 ----------------------
3025 1. Added Markus Oberhumer's little patches for C++.
3027 2. Literal strings longer than 255 characters were broken.
3030 Version 1.05 23-Dec-97
3031 ----------------------
3033 1. Negated character classes containing more than one character were failing if
3034 PCRE_CASELESS was set at run time.
3037 Version 1.04 19-Dec-97
3038 ----------------------
3040 1. Corrected the man page, where some "const" qualifiers had been omitted.
3042 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3045 3. Fixed memory leak which occurred when a regex with back references was
3046 matched with an offsets vector that wasn't big enough. The temporary memory
3047 that is used in this case wasn't being freed if the match failed.
3049 4. Tidied pcretest to ensure it frees memory that it gets.
3051 5. Temporary memory was being obtained in the case where the passed offsets
3052 vector was exactly big enough.
3054 6. Corrected definition of offsetof() from change 5 below.
3056 7. I had screwed up change 6 below and broken the rules for the use of
3057 setjmp(). Now fixed.
3060 Version 1.03 18-Dec-97
3061 ----------------------
3063 1. A erroneous regex with a missing opening parenthesis was correctly
3064 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3067 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3068 it was reported that one broken compiler failed on the former because "code" is
3069 also an independent variable.
3071 3. The erroneous regex a[]b caused an array overrun reference.
3073 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3074 fail on data ending with that character. (It was going on too far, and checking
3075 the next character, typically a binary zero.) This was specific to the
3076 optimized code for single-character negative classes.
3078 5. Added a contributed patch from the TIN world which does the following:
3080 + Add an undef for memmove, in case the the system defines a macro for it.
3082 + Add a definition of offsetof(), in case there isn't one. (I don't know
3083 the reason behind this - offsetof() is part of the ANSI standard - but
3086 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3087 most of the places where whitespace preceded '#'. I have given up and
3088 allowed the remaining 2 cases to be at the margin.
3090 + Rename some variables in pcre to eliminate shadowing. This seems very
3091 pedantic, but does no harm, of course.
3093 6. Moved the call to setjmp() into its own function, to get rid of warnings
3094 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3096 7. Constructs such as \d{8,} were compiling into the equivalent of
3097 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3098 outcome, but in this particular case used more store than had been allocated,
3099 which caused the bug to be discovered because it threw up an internal error.
3101 8. The debugging code in both pcre and pcretest for outputting the compiled
3102 form of a regex was going wrong in the case of back references followed by
3103 curly-bracketed repeats.
3106 Version 1.02 12-Dec-97
3107 ----------------------
3109 1. Typos in pcre.3 and comments in the source fixed.
3111 2. Applied a contributed patch to get rid of places where it used to remove
3112 'const' from variables, and fixed some signed/unsigned and uninitialized
3115 3. Added the "runtest" target to Makefile.
3117 4. Set default compiler flag to -O2 rather than just -O.
3120 Version 1.01 19-Nov-97
3121 ----------------------
3123 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3124 like /([ab]*)*/, that is, for classes with more than one character in them.
3126 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3127 as /((?>a*))*/ (a PCRE_EXTRA facility).
3130 Version 1.00 18-Nov-97
3131 ----------------------
3133 1. Added compile-time macros to support systems such as SunOS4 which don't have
3134 memmove() or strerror() but have other things that can be used instead.
3136 2. Arranged that "make clean" removes the executables.
3139 Version 0.99 27-Oct-97
3140 ----------------------
3142 1. Fixed bug in code for optimizing classes with only one character. It was
3143 initializing a 32-byte map regardless, which could cause it to run off the end
3144 of the memory it had got.
3146 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3149 Version 0.98 22-Oct-97
3150 ----------------------
3152 1. Fixed bug in code for handling temporary memory usage when there are more
3153 back references than supplied space in the ovector. This could cause segfaults.
3156 Version 0.97 21-Oct-97
3157 ----------------------
3159 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3161 2. Optimized negated single characters not to use a bit map.
3163 3. Brought error texts together as macro definitions; clarified some of them;
3164 fixed one that was wrong - it said "range out of order" when it meant "invalid
3167 4. Changed some char * arguments to const char *.
3169 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3171 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3175 Version 0.96 16-Oct-97
3176 ----------------------
3178 1. Added a simple "pgrep" utility to the distribution.
3180 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3181 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3182 where "ddd" means "one or more decimal digits".
3184 3. Fixed serious bug. If a pattern had a back reference, but the call to
3185 pcre_exec() didn't supply a large enough ovector to record the related
3186 identifying subpattern, the match always failed. PCRE now remembers the number
3187 of the largest back reference, and gets some temporary memory in which to save
3188 the offsets during matching if necessary, in order to ensure that
3189 backreferences always work.
3191 4. Increased the compatibility with Perl in a number of ways:
3193 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3194 to request this handling. The option can be set at compile or exec time.
3196 (b) $ matches before a terminating newline by default; an option
3197 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3198 mode). The option can be set at compile or exec time.
3200 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3201 the same as Perl's. If the decimal number it represents is less than 10
3202 or there aren't that many previous left capturing parentheses, an octal
3203 escape is read. Inside a character class, it's always an octal escape,
3204 even if it is a single digit.
3206 (d) An escaped but undefined alphabetic character is taken as a literal,
3207 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3210 (e) {0} is now permitted. (The previous item is removed from the compiled
3213 5. Changed all the names of code files so that the basic parts are no longer
3214 than 10 characters, and abolished the teeny "globals.c" file.
3216 6. Changed the handling of character classes; they are now done with a 32-byte
3219 7. Added the -d and /D options to pcretest to make it possible to look at the
3220 internals of compilation without having to recompile pcre.
3223 Version 0.95 23-Sep-97
3224 ----------------------
3226 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3227 \x20 at the start of a run of normal characters. These were being treated as
3228 real characters, instead of the source characters being re-checked.
3231 Version 0.94 18-Sep-97
3232 ----------------------
3234 1. The functions are now thread-safe, with the caveat that the global variables
3235 containing pointers to malloc() and free() or alternative functions are the
3236 same for all threads.
3238 2. Get pcre_study() to generate a bitmap of initial characters for non-
3239 anchored patterns when this is possible, and use it if passed to pcre_exec().
3242 Version 0.93 15-Sep-97
3243 ----------------------
3245 1. /(b)|(:+)/ was computing an incorrect first character.
3247 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3248 but not actually doing anything yet.
3250 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3251 as Perl does (e.g. [-az] or [az-]).
3253 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3254 all possible positions.
3256 5. Split up into different modules to avoid including unneeded functions in a
3257 compiled binary. However, compile and exec are still in one module. The "study"
3258 function is split off.
3260 6. The character tables are now in a separate module whose source is generated
3261 by an auxiliary program - but can then be edited by hand if required. There are
3262 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3263 toupper() in the code.
3265 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3266 make them global. Abolish the function for setting them, as the caller can now
3270 Version 0.92 11-Sep-97
3271 ----------------------
3273 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3274 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3276 2. Caseless matching was not working in character classes if the characters in
3277 the pattern were in upper case.
3279 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3281 4. Make PCRE_ANCHORED public and accept as a compile option.
3283 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3284 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3287 6. Give an error if bad option bits passed at compile or run time.
3289 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3290 pcretest to cause it to pass that flag.
3292 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3293 options, and the first character, if set.
3295 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3298 Version 0.91 10-Sep-97
3299 ----------------------
3301 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3302 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3304 2. PCRE was looping on encountering an indefinitely repeated back reference to
3305 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3306 Perl does - treats the match as successful.