ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 270 - (show annotations)
Fri Nov 16 19:55:16 2007 UTC (7 years, 10 months ago) by ph10
File size: 143418 byte(s)
Error occurred while calculating annotation data.
Diagnose conditional numerical reference to a non-existent subpattern.
1 ChangeLog for PCRE
2 ------------------
4 Version 7.5 12-Nov-07
5 ---------------------
7 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
8 values in parens when parsing an RE using the C++ wrapper."
10 2. Negative specials like \S did not work in character classes in UTF-8 mode.
11 Characters greater than 255 were excluded from the class instead of being
12 included.
14 3. The same bug as (2) above applied to negated POSIX classes such as
15 [:^space:].
17 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
18 defined or documented. It seems to have been a typo for PCRE_STATIC, so
19 I have changed it.
21 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
22 first named subpattern) and a construct such as (?&a) would reference the
23 first named subpattern whose name started with "a" (in other words, the
24 length check was missing).
26 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
27 this is right because (?- can be followed by option settings as well as by
28 digits. I have, however, made the messages clearer.
30 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
31 than the number used in the conditional) now cause a compile-time error.
32 This is actually not compatible with Perl, which accepts such patterns, but
33 treats the conditional as always being FALSE (as PCRE used to), but it
34 seems to me that giving a diagnostic is better.
37 Version 7.4 21-Sep-07
38 ---------------------
40 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
41 means that a class such as [\s] counted as "explicit reference to CR or
42 LF". That isn't really right - the whole point of the change was to try to
43 help when there was an actual mention of one of the two characters. So now
44 the change happens only if \r or \n (or a literal CR or LF) character is
45 encountered.
47 2. The 32-bit options word was also used for 6 internal flags, but the numbers
48 of both had grown to the point where there were only 3 bits left.
49 Fortunately, there was spare space in the data structure, and so I have
50 moved the internal flags into a new 16-bit field to free up more option
51 bits.
53 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
54 but did not set the internal JCHANGED flag - either of these is enough to
55 control the way the "get" function works - but the PCRE_INFO_JCHANGED
56 facility is supposed to tell if (?J) was ever used, so now (?J) at the
57 start sets both bits.
59 4. Added options (at build time, compile time, exec time) to change \R from
60 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
62 5. doc/pcresyntax.html was missing from the distribution.
64 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
65 compatibility, even though it is no longer used.
67 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
68 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
69 windows.h file is present (where different names are used). [This was
70 reversed later after testing - see 16 below.]
72 8. Changed all #include <config.h> to #include "config.h". There were also
73 some further <pcre.h> cases that I changed to "pcre.h".
75 9. When pcregrep was used with the --colour option, it missed the line ending
76 sequence off the lines that it output.
78 10. It was pointed out to me that arrays of string pointers cause lots of
79 relocations when a shared library is dynamically loaded. A technique of
80 using a single long string with a table of offsets can drastically reduce
81 these. I have refactored PCRE in four places to do this. The result is
82 dramatic:
84 Originally: 290
85 After changing UCP table: 187
86 After changing error message table: 43
87 After changing table of "verbs" 36
88 After changing table of Posix names 22
90 Thanks to the folks working on Gregex for glib for this insight.
92 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
93 unicode-properties was also set.
95 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
97 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
98 checked only for CRLF.
100 14. Added casts to pcretest.c to avoid compiler warnings.
102 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
104 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
105 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
106 entirely. This removes changes made in 7 above.
108 17. The CMake files have been updated, and there is now more information about
109 building with CMake in the NON-UNIX-USE document.
112 Version 7.3 28-Aug-07
113 ---------------------
115 1. In the rejigging of the build system that eventually resulted in 7.1, the
116 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
117 brackets there is not right, since it causes compilers to look for an
118 installed pcre.h, not the version that is in the source that is being
119 compiled (which of course may be different). I have changed it back to:
121 #include "pcre.h"
123 I have a vague recollection that the change was concerned with compiling in
124 different directories, but in the new build system, that is taken care of
125 by the VPATH setting the Makefile.
127 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
128 when the subject happened to end in the byte 0x85 (e.g. if the last
129 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
130 characters but of course it shouldn't be taken as a newline when it is part
131 of another character. The bug was that, for an unlimited repeat of . in
132 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
133 characters when looking for a newline.
135 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
137 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
138 in debug output.
140 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
141 long printing in the pcrecpp unittest when running under MinGW.
143 6. ESC_K was left out of the EBCDIC table.
145 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
146 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
147 limit also applies to "virtual nesting" when a pattern is recursive, and in
148 this case 1000 isn't so big. I have been able to remove this limit at the
149 expense of backing off one optimization in certain circumstances. Normally,
150 when pcre_exec() would call its internal match() function recursively and
151 immediately return the result unconditionally, it uses a "tail recursion"
152 feature to save stack. However, when a subpattern that can match an empty
153 string has an unlimited repetition quantifier, it no longer makes this
154 optimization. That gives it a stack frame in which to save the data for
155 checking that an empty string has been matched. Previously this was taken
156 from the 1000-entry workspace that had been reserved. So now there is no
157 explicit limit, but more stack is used.
159 8. Applied Daniel's patches to solve problems with the import/export magic
160 syntax that is required for Windows, and which was going wrong for the
161 pcreposix and pcrecpp parts of the library. These were overlooked when this
162 problem was solved for the main library.
164 9. There were some crude static tests to avoid integer overflow when computing
165 the size of patterns that contain repeated groups with explicit upper
166 limits. As the maximum quantifier is 65535, the maximum group length was
167 set at 30,000 so that the product of these two numbers did not overflow a
168 32-bit integer. However, it turns out that people want to use groups that
169 are longer than 30,000 bytes (though not repeat them that many times).
170 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
171 made it possible to implement the integer overflow checks in a much more
172 dynamic way, which I have now done. The artificial limitation on group
173 length has been removed - we now have only the limit on the total length of
174 the compiled pattern, which depends on the LINK_SIZE setting.
176 10. Fixed a bug in the documentation for get/copy named substring when
177 duplicate names are permitted. If none of the named substrings are set, the
178 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
179 empty string.
181 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
182 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
183 because the ] is interpreted as the first data character and the
184 terminating ] is not found. PCRE has been made compatible with Perl in this
185 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
186 cause memory overwriting.
188 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
189 string has been matched (to stop an infinite loop). It was not recognizing
190 a conditional subpattern that could match an empty string if that
191 subpattern was within another subpattern. For example, it looped when
192 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
193 condition was not nested. This bug has been fixed.
195 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
196 past the start of the subject in the presence of bytes with the top bit
197 set, for example "\x8aBCD".
199 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
200 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
202 14. Optimized (?!) to (*FAIL).
204 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
205 This restricts code points to be within the range 0 to 0x10FFFF, excluding
206 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
207 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
208 does: it's just the validity check that is more restrictive.
210 16. Inserted checks for integer overflows during escape sequence (backslash)
211 processing, and also fixed erroneous offset values for syntax errors during
212 backslash processing.
214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
215 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
217 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
218 caused an overrun.
220 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
221 something other than just ASCII characters) inside a group that had an
222 unlimited repeat caused a loop at compile time (while checking to see
223 whether the group could match an empty string).
225 20. Debugging a pattern containing \p or \P could cause a crash. For example,
226 [\P{Any}] did so. (Error in the code for printing property names.)
228 21. An orphan \E inside a character class could cause a crash.
230 22. A repeated capturing bracket such as (A)? could cause a wild memory
231 reference during compilation.
233 23. There are several functions in pcre_compile() that scan along a compiled
234 expression for various reasons (e.g. to see if it's fixed length for look
235 behind). There were bugs in these functions when a repeated \p or \P was
236 present in the pattern. These operators have additional parameters compared
237 with \d, etc, and these were not being taken into account when moving along
238 the compiled data. Specifically:
240 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
241 length.
243 (b) An item such as \pL+ within a repeated group could cause crashes or
244 loops.
246 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
247 "reference to non-existent subpattern" error.
249 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
251 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
252 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
254 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
256 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
257 character were causing crashes (broken optimization).
259 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
260 \p or \P) caused a compile-time loop.
262 28. More problems have arisen in unanchored patterns when CRLF is a valid line
263 break. For example, the unstudied pattern [\r\n]A does not match the string
264 "\r\nA" because change 7.0/46 below moves the current point on by two
265 characters after failing to match at the start. However, the pattern \nA
266 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
267 the same is true. There doesn't seem any very clean way out of this, but
268 what I have chosen to do makes the common cases work: PCRE now takes note
269 of whether there can be an explicit match for \r or \n anywhere in the
270 pattern, and if so, 7.0/46 no longer applies. As part of this change,
271 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
272 pattern has explicit CR or LF references.
274 29. Added (*CR) etc for changing newline setting at start of pattern.
277 Version 7.2 19-Jun-07
278 ---------------------
280 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
281 which is apparently normally available under Windows.
283 2. Re-jig the pcregrep tests with different newline settings in an attempt
284 to make them independent of the local environment's newline setting.
286 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
288 4. Some of the "internals" tests were previously cut out when the link size
289 was not 2, because the output contained actual offsets. The recent new
290 "Z" feature of pcretest means that these can be cut out, making the tests
291 usable with all link sizes.
293 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
294 stack recursion. This gives a massive performance boost under BSD, but just
295 a small improvement under Linux. However, it saves one field in the frame
296 in all cases.
298 6. Added more features from the forthcoming Perl 5.10:
300 (a) (?-n) (where n is a string of digits) is a relative subroutine or
301 recursion call. It refers to the nth most recently opened parentheses.
303 (b) (?+n) is also a relative subroutine call; it refers to the nth next
304 to be opened parentheses.
306 (c) Conditions that refer to capturing parentheses can be specified
307 relatively, for example, (?(-2)... or (?(+3)...
309 (d) \K resets the start of the current match so that everything before
310 is not part of it.
312 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
314 (f) \g{name} is another synonym - part of Perl 5.10's unification of
315 reference syntax.
317 (g) (?| introduces a group in which the numbering of parentheses in each
318 alternative starts with the same number.
320 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
322 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
325 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
326 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
327 for detecting groups that can match an empty string.
329 9. A pattern with a very large number of alternatives (more than several
330 hundred) was running out of internal workspace during the pre-compile
331 phase, where pcre_compile() figures out how much memory will be needed. A
332 bit of new cunning has reduced the workspace needed for groups with
333 alternatives. The 1000-alternative test pattern now uses 12 bytes of
334 workspace instead of running out of the 4096 that are available.
336 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
338 11. Applied patch from Google to remove an optimization that didn't quite work.
339 The report of the bug said:
341 pcrecpp::RE("a*").FullMatch("aaa") matches, while
342 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
343 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
345 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
346 it matched the wrong number of bytes.
349 Version 7.1 24-Apr-07
350 ---------------------
352 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
353 that is more "standard", making use of automake and other Autotools. There
354 is some re-arrangement of the files and adjustment of comments consequent
355 on this.
357 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
358 for recursive directory scanning broke on some systems because the files
359 are not scanned in any specific order and on different systems the order
360 was different. A call to "sort" has been inserted into RunGrepTest for the
361 approprate test as a short-term fix. In the longer term there may be an
362 alternative.
364 3. I had an email from Eric Raymond about problems translating some of PCRE's
365 man pages to HTML (despite the fact that I distribute HTML pages, some
366 people do their own conversions for various reasons). The problems
367 concerned the use of low-level troff macros .br and .in. I have therefore
368 removed all such uses from the man pages (some were redundant, some could
369 be replaced by .nf/.fi pairs). The 132html script that I use to generate
370 HTML has been updated to handle .nf/.fi and to complain if it encounters
371 .br or .in.
373 4. Updated comments in configure.ac that get placed in config.h.in and also
374 arranged for config.h to be included in the distribution, with the name
375 config.h.generic, for the benefit of those who have to compile without
376 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
378 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
379 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
380 makevp.bat for the current PCRE, using the additional files
381 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
383 6. A Windows user reported a minor discrepancy with test 2, which turned out
384 to be caused by a trailing space on an input line that had got lost in his
385 copy. The trailing space was an accident, so I've just removed it.
387 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
388 that is needed.
390 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
391 as "const" (a) because they are and (b) because it helps the PHP
392 maintainers who have recently made a script to detect big data structures
393 in the php code that should be moved to the .rodata section. I remembered
394 to update Builducptable as well, so it won't revert if ucptable.h is ever
395 re-created.
397 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
398 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
399 order to be able to cut out the UTF-8 tables in the latter when UTF-8
400 support is not required. This saves 1.5-2K of code, which is important in
401 some applications.
403 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
404 so as not to refer to the tables, even though these functions will never be
405 called when UTF-8 support is disabled. Otherwise there are problems with a
406 shared library.
408 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
410 (a) It was defining its arguments as char * instead of void *.
412 (b) It was assuming that all moves were upwards in memory; this was true
413 a long time ago when I wrote it, but is no longer the case.
415 The emulated memove() is provided for those environments that have neither
416 memmove() nor bcopy(). I didn't think anyone used it these days, but that
417 is clearly not the case, as these two bugs were recently reported.
419 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
420 and Detrail to create the HTML documentation, the .txt form of the man
421 pages, and it removes trailing spaces from listed files. It also creates
422 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
423 case, it wraps all the #defines with #ifndefs. This script should be run
424 before "make dist".
426 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
427 with Unicode property support.
429 (a) For a maximizing quantifier, if the two different cases of the
430 character were of different lengths in their UTF-8 codings (there are
431 some cases like this - I found 11), and the matching function had to
432 back up over a mixture of the two cases, it incorrectly assumed they
433 were both the same length.
435 (b) When PCRE was configured to use the heap rather than the stack for
436 recursion during matching, it was not correctly preserving the data for
437 the other case of a UTF-8 character when checking ahead for a match
438 while processing a minimizing repeat. If the check also involved
439 matching a wide character, but failed, corruption could cause an
440 erroneous result when trying to check for a repeat of the original
441 character.
443 13. Some tidying changes to the testing mechanism:
445 (a) The RunTest script now detects the internal link size and whether there
446 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
447 values substituted by "configure". (The RunGrepTest script already did
448 this for UTF-8.) The configure.ac script no longer substitutes the
449 relevant variables.
451 (b) The debugging options /B and /D in pcretest show the compiled bytecode
452 with length and offset values. This means that the output is different
453 for different internal link sizes. Test 2 is skipped for link sizes
454 other than 2 because of this, bypassing the problem. Unfortunately,
455 there was also a test in test 3 (the locale tests) that used /B and
456 failed for link sizes other than 2. Rather than cut the whole test out,
457 I have added a new /Z option to pcretest that replaces the length and
458 offset values with spaces. This is now used to make test 3 independent
459 of link size. (Test 2 will be tidied up later.)
461 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
462 segmentation fault instead of returning the appropriate error message.
464 15. In multiline mode when the newline sequence was set to "any", the pattern
465 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
466 This doesn't seem right; it now treats the CRLF combination as the line
467 ending, and so does not match in that case. It's only a pattern such as ^$
468 that would hit this one: something like ^ABC$ would have failed after \r
469 and then tried again after \r\n.
471 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
472 in an attempt to make files that differ only in their line terminators
473 compare equal. This works on Linux.
475 17. Under certain error circumstances pcregrep might try to free random memory
476 as it exited. This is now fixed, thanks to valgrind.
478 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
479 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
480 was because its rules for how to advance for /g after matching an empty
481 string at the end of a line did not allow for this case. They now check for
482 it specially.
484 20. pcretest is supposed to handle patterns and data of any length, by
485 extending its buffers when necessary. It was getting this wrong when the
486 buffer for a data line had to be extended.
488 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
489 CRLF as a newline sequence.
491 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
492 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
493 I have nevertheless tidied it up.
495 23. Added some casts to kill warnings from HP-UX ia64 compiler.
497 24. Added a man page for pcre-config.
500 Version 7.0 19-Dec-06
501 ---------------------
503 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
504 moving to gcc 4.1.1.
506 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
507 sys/time.h, which is documented as needed for this function. It doesn't
508 seem to matter on Linux, but it showed up on some releases of OS X.
510 3. It seems that there are systems where bytes whose values are greater than
511 127 match isprint() in the "C" locale. The "C" locale should be the
512 default when a C program starts up. In most systems, only ASCII printing
513 characters match isprint(). This difference caused the output from pcretest
514 to vary, making some of the tests fail. I have changed pcretest so that:
516 (a) When it is outputting text in the compiled version of a pattern, bytes
517 other than 32-126 are always shown as hex escapes.
519 (b) When it is outputting text that is a matched part of a subject string,
520 it does the same, unless a different locale has been set for the match
521 (using the /L modifier). In this case, it uses isprint() to decide.
523 4. Fixed a major bug that caused incorrect computation of the amount of memory
524 required for a compiled pattern when options that changed within the
525 pattern affected the logic of the preliminary scan that determines the
526 length. The relevant options are -x, and -i in UTF-8 mode. The result was
527 that the computed length was too small. The symptoms of this bug were
528 either the PCRE error "internal error: code overflow" from pcre_compile(),
529 or a glibc crash with a message such as "pcretest: free(): invalid next
530 size (fast)". Examples of patterns that provoked this bug (shown in
531 pcretest format) are:
533 /(?-x: )/x
534 /(?x)(?-x: \s*#\s*)/
535 /((?i)[\x{c0}])/8
536 /(?i:[\x{c0}])/8
538 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
539 is now done differently.
541 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
542 wrapper classes; (b) implement a new function in the C++ scanner that is
543 more efficient than the old way of doing things because it avoids levels of
544 recursion in the regex matching; (c) add a paragraph to the documentation
545 for the FullMatch() function.
547 6. The escape sequence \n was being treated as whatever was defined as
548 "newline". Not only was this contrary to the documentation, which states
549 that \n is character 10 (hex 0A), but it also went horribly wrong when
550 "newline" was defined as CRLF. This has been fixed.
552 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
553 was being set to -1 for the "end of line" case (supposedly a value that no
554 character can have). Though this value is never used (the check for end of
555 line is "zero bytes in current character"), it caused compiler complaints.
556 I've changed it to 0xffffffff.
558 8. In pcre_version.c, the version string was being built by a sequence of
559 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
560 string (as it is for production releases) called a macro with an empty
561 argument. The C standard says the result of this is undefined. The gcc
562 compiler treats it as an empty string (which was what was wanted) but it is
563 reported that Visual C gives an error. The source has been hacked around to
564 avoid this problem.
566 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
567 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
568 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
569 of them did).
571 10. Originally, pcretest opened its input and output without "b"; then I was
572 told that "b" was needed in some environments, so it was added for release
573 5.0 to both the input and output. (It makes no difference on Unix-like
574 systems.) Later I was told that it is wrong for the input on Windows. I've
575 now abstracted the modes into two macros, to make it easier to fiddle with
576 them, and removed "b" from the input mode under Windows.
578 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
580 12. Added -help and --help to pcretest as an official way of being reminded
581 of the options.
583 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
584 and pcrecpp.cc because they annoy compilers at high warning levels.
586 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
588 15. Fixed an occurrence of == in configure.ac that should have been = (shell
589 scripts are not C programs :-) and which was not noticed because it works
590 on Linux.
592 16. pcretest is supposed to handle any length of pattern and data line (as one
593 line or as a continued sequence of lines) by extending its input buffer if
594 necessary. This feature was broken for very long pattern lines, leading to
595 a string of junk being passed to pcre_compile() if the pattern was longer
596 than about 50K.
598 17. I have done a major re-factoring of the way pcre_compile() computes the
599 amount of memory needed for a compiled pattern. Previously, there was code
600 that made a preliminary scan of the pattern in order to do this. That was
601 OK when PCRE was new, but as the facilities have expanded, it has become
602 harder and harder to keep it in step with the real compile phase, and there
603 have been a number of bugs (see for example, 4 above). I have now found a
604 cunning way of running the real compile function in a "fake" mode that
605 enables it to compute how much memory it would need, while actually only
606 ever using a few hundred bytes of working memory and without too many
607 tests of the mode. This should make future maintenance and development
608 easier. A side effect of this work is that the limit of 200 on the nesting
609 depth of parentheses has been removed (though this was never a serious
610 limitation, I suspect). However, there is a downside: pcre_compile() now
611 runs more slowly than before (30% or more, depending on the pattern). I
612 hope this isn't a big issue. There is no effect on runtime performance.
614 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
615 newline (only possible for the last line of a file) and it was a
616 pattern that set a locale (followed by /Lsomething), pcretest crashed.
618 19. Added additional timing features to pcretest. (1) The -tm option now times
619 matching only, not compiling. (2) Both -t and -tm can be followed, as a
620 separate command line item, by a number that specifies the number of
621 repeats to use when timing. The default is 50000; this gives better
622 precision, but takes uncomfortably long for very large patterns.
624 20. Extended pcre_study() to be more clever in cases where a branch of a
625 subpattern has no definite first character. For example, (a*|b*)[cd] would
626 previously give no result from pcre_study(). Now it recognizes that the
627 first character must be a, b, c, or d.
629 21. There was an incorrect error "recursive call could loop indefinitely" if
630 a subpattern (or the entire pattern) that was being tested for matching an
631 empty string contained only one non-empty item after a nested subpattern.
632 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
633 incorrectly, because the \d was being skipped in the check.
635 22. The pcretest program now has a new pattern option /B and a command line
636 option -b, which is equivalent to adding /B to every pattern. This causes
637 it to show the compiled bytecode, without the additional information that
638 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
639 is the same as /B/I).
641 23. A new optimization is now able automatically to treat some sequences such
642 as a*b as a*+b. More specifically, if something simple (such as a character
643 or a simple class like \d) has an unlimited quantifier, and is followed by
644 something that cannot possibly match the quantified thing, the quantifier
645 is automatically "possessified".
647 24. A recursive reference to a subpattern whose number was greater than 39
648 went wrong under certain circumstances in UTF-8 mode. This bug could also
649 have affected the operation of pcre_study().
651 25. Realized that a little bit of performance could be had by replacing
652 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
654 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
656 27. Possessive quantifiers such as a++ were previously implemented by turning
657 them into atomic groups such as ($>a+). Now they have their own opcodes,
658 which improves performance. This includes the automatically created ones
659 from 23 above.
661 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
662 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
663 the first matched character to be a colon. This applied both to named and
664 numbered groups.
666 29. The ucpinternal.h header file was missing its idempotency #ifdef.
668 30. I was sent a "project" file called libpcre.a.dev which I understand makes
669 building PCRE on Windows easier, so I have included it in the distribution.
671 31. There is now a check in pcretest against a ridiculously large number being
672 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
673 loop, the loop is abandoned.
675 32. Forward references to subpatterns in conditions such as (?(2)...) where
676 subpattern 2 is defined later cause pcre_compile() to search forwards in
677 the pattern for the relevant set of parentheses. This search went wrong
678 when there were unescaped parentheses in a character class, parentheses
679 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
681 33. "Subroutine" calls and backreferences were previously restricted to
682 referencing subpatterns earlier in the regex. This restriction has now
683 been removed.
685 34. Added a number of extra features that are going to be in Perl 5.10. On the
686 whole, these are just syntactic alternatives for features that PCRE had
687 previously implemented using the Python syntax or my own invention. The
688 other formats are all retained for compatibility.
690 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
691 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
692 also .NET compatible.
694 (b) A recursion or subroutine call to a named group can now be defined as
695 (?&name) as well as (?P>name).
697 (c) A backreference to a named group can now be defined as \k<name> or
698 \k'name' as well as (?P=name). The new forms, as well as being in Perl
699 5.10, are also .NET compatible.
701 (d) A conditional reference to a named group can now use the syntax
702 (?(<name>) or (?('name') as well as (?(name).
704 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
705 groups (named and numbered) that are never evaluated inline, but can be
706 called as "subroutines" from elsewhere. In effect, the DEFINE condition
707 is always false. There may be only one alternative in such a group.
709 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
710 as the simple (?(R). The condition is true only if the most recent
711 recursion is that of the given number or name. It does not search out
712 through the entire recursion stack.
714 (g) The escape \gN or \g{N} has been added, where N is a positive or
715 negative number, specifying an absolute or relative reference.
717 35. Tidied to get rid of some further signed/unsigned compiler warnings and
718 some "unreachable code" warnings.
720 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
721 things, this adds five new scripts.
723 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
724 There were also incompatibilities regarding the handling of \Q..\E inside
725 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
726 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
728 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
729 matches an empty string, and forcibly breaks the loop. There were bugs in
730 this code in non-simple cases. For a pattern such as ^(a()*)* matched
731 against aaaa the result was just "a" rather than "aaaa", for example. Two
732 separate and independent bugs (that affected different cases) have been
733 fixed.
735 39. Refactored the code to abolish the use of different opcodes for small
736 capturing bracket numbers. This is a tidy that I avoided doing when I
737 removed the limit on the number of capturing brackets for 3.5 back in 2001.
738 The new approach is not only tidier, it makes it possible to reduce the
739 memory needed to fix the previous bug (38).
741 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
742 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
743 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
744 mode.
746 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
747 report.
749 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
750 copy construction and assignment in the C++ wrapper.
752 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
753 couple of bugs that could have given wrong results in the "--newline=crlf"
754 case.
756 44. Added a number of casts and did some reorganization of signed/unsigned int
757 variables following suggestions from Dair Grant. Also renamed the variable
758 "this" as "item" because it is a C++ keyword.
760 45. Arranged for dftables to add
762 #include "pcre_internal.h"
764 to pcre_chartables.c because without it, gcc 4.x may remove the array
765 definition from the final binary if PCRE is built into a static library and
766 dead code stripping is activated.
768 46. For an unanchored pattern, if a match attempt fails at the start of a
769 newline sequence, and the newline setting is CRLF or ANY, and the next two
770 characters are CRLF, advance by two characters instead of one.
773 Version 6.7 04-Jul-06
774 ---------------------
776 1. In order to handle tests when input lines are enormously long, pcretest has
777 been re-factored so that it automatically extends its buffers when
778 necessary. The code is crude, but this _is_ just a test program. The
779 default size has been increased from 32K to 50K.
781 2. The code in pcre_study() was using the value of the re argument before
782 testing it for NULL. (Of course, in any sensible call of the function, it
783 won't be NULL.)
785 3. The memmove() emulation function in pcre_internal.h, which is used on
786 systems that lack both memmove() and bcopy() - that is, hardly ever -
787 was missing a "static" storage class specifier.
789 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
790 containing an extended class (one that cannot be represented by a bitmap
791 because it contains high-valued characters or Unicode property items, e.g.
792 [\pZ]). Almost always one would set UTF-8 mode when processing such a
793 pattern, but PCRE should not loop if you do not (it no longer does).
794 [Detail: two cases were found: (a) a repeated subpattern containing an
795 extended class; (b) a recursive reference to a subpattern that followed a
796 previous extended class. It wasn't skipping over the extended class
797 correctly when UTF-8 mode was not set.]
799 5. A negated single-character class was not being recognized as fixed-length
800 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
801 compile error "lookbehind assertion is not fixed length".
803 6. The RunPerlTest auxiliary script was showing an unexpected difference
804 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
805 write a Perl script that can interpret lines of an input file either as
806 byte characters or as UTF-8, which is what "perltest" was being required to
807 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
808 can't do is switch easily at run time between having the "use utf8;" pragma
809 or not. In the end, I fudged it by using the RunPerlTest script to insert
810 "use utf8;" explicitly for the UTF-8 tests.
812 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
813 the end of the subject string, contrary to the documentation and to what
814 Perl does. This was true of both matching functions. Now it matches only at
815 the start of the subject and immediately after *internal* newlines.
817 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
818 a pointer to an int instead of a pointer to an unsigned long int. This
819 caused problems on 64-bit systems.
821 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
822 instance of the 'standard' template library not being so standard".
824 10. There was no check on the number of named subpatterns nor the maximum
825 length of a subpattern name. The product of these values is used to compute
826 the size of the memory block for a compiled pattern. By supplying a very
827 long subpattern name and a large number of named subpatterns, the size
828 computation could be caused to overflow. This is now prevented by limiting
829 the length of names to 32 characters, and the number of named subpatterns
830 to 10,000.
832 11. Subpatterns that are repeated with specific counts have to be replicated in
833 the compiled pattern. The size of memory for this was computed from the
834 length of the subpattern and the repeat count. The latter is limited to
835 65535, but there was no limit on the former, meaning that integer overflow
836 could in principle occur. The compiled length of a repeated subpattern is
837 now limited to 30,000 bytes in order to prevent this.
839 12. Added the optional facility to have named substrings with the same name.
841 13. Added the ability to use a named substring as a condition, using the
842 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
843 are numbers (not recommended). Forward references are permitted.
845 14. Added forward references in named backreferences (if you see what I mean).
847 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
848 pattern could run off the end of the subject. For example, the pattern
849 "(?s)(.{1,5})"8 did this with the subject "ab".
851 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
852 PCRE_CASELESS was set when matching characters that were quantified with ?
853 or *.
855 17. A character class other than a single negated character that had a minimum
856 but no maximum quantifier - for example [ab]{6,} - was not handled
857 correctly by pce_dfa_exec(). It would match only one character.
859 18. A valid (though odd) pattern that looked like a POSIX character
860 class but used an invalid character after [ (for example [[,abc,]]) caused
861 pcre_compile() to give the error "Failed: internal error: code overflow" or
862 in some cases to crash with a glibc free() error. This could even happen if
863 the pattern terminated after [[ but there just happened to be a sequence of
864 letters, a binary zero, and a closing ] in the memory that followed.
866 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
867 over the years. Originally (before any Unicode support), just the bottom 8
868 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
869 output from "man perlunicode" includes this:
871 The regular expression compiler produces polymorphic opcodes. That
872 is, the pattern adapts to the data and automatically switches to
873 the Unicode character scheme when presented with Unicode data--or
874 instead uses a traditional byte scheme when presented with byte
875 data.
877 Sadly, a wide octal escape does not cause a switch, and in a string with
878 no other multibyte characters, these octal escapes are treated as before.
879 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
880 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
881 Unicode string.
883 I have not perpetrated such confusion in PCRE. Up till now, it took just
884 the bottom 8 bits, as in old Perl. I have now made octal escapes with
885 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
886 translate to the appropriate multibyte character.
888 29. Applied some refactoring to reduce the number of warnings from Microsoft
889 and Borland compilers. This has included removing the fudge introduced
890 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
891 a warning about an unused variable.
893 21. PCRE has not included VT (character 0x0b) in the set of whitespace
894 characters since release 4.0, because Perl (from release 5.004) does not.
895 [Or at least, is documented not to: some releases seem to be in conflict
896 with the documentation.] However, when a pattern was studied with
897 pcre_study() and all its branches started with \s, PCRE still included VT
898 as a possible starting character. Of course, this did no harm; it just
899 caused an unnecessary match attempt.
901 22. Removed a now-redundant internal flag bit that recorded the fact that case
902 dependency changed within the pattern. This was once needed for "required
903 byte" processing, but is no longer used. This recovers a now-scarce options
904 bit. Also moved the least significant internal flag bit to the most-
905 significant bit of the word, which was not previously used (hangover from
906 the days when it was an int rather than a uint) to free up another bit for
907 the future.
909 23. Added support for CRLF line endings as well as CR and LF. As well as the
910 default being selectable at build time, it can now be changed at runtime
911 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
912 specify that it is scanning data with non-default line endings.
914 24. Changed the definition of CXXLINK to make it agree with the definition of
915 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
917 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
918 recursions. This makes a big different to stack usage for some patterns.
920 26. If a subpattern containing a named recursion or subroutine reference such
921 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
922 the space required for the compiled pattern went wrong and gave too small a
923 value. Depending on the environment, this could lead to "Failed: internal
924 error: code overflow at offset 49" or "glibc detected double free or
925 corruption" errors.
927 27. Applied patches from Google (a) to support the new newline modes and (b) to
928 advance over multibyte UTF-8 characters in GlobalReplace.
930 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
931 difference for some implementation of PCRE in some Windows version.
933 29. Added some extra testing facilities to pcretest:
935 \q<number> in a data line sets the "match limit" value
936 \Q<number> in a data line sets the "match recursion limt" value
937 -S <number> sets the stack size, where <number> is in megabytes
939 The -S option isn't available for Windows.
942 Version 6.6 06-Feb-06
943 ---------------------
945 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
946 in pcreposix.h. I have copied the definition from pcre.h.
948 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
949 because pcre.h is no longer a built file.
951 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
952 not normally included in the compiled code.
955 Version 6.5 01-Feb-06
956 ---------------------
958 1. When using the partial match feature with pcre_dfa_exec(), it was not
959 anchoring the second and subsequent partial matches at the new starting
960 point. This could lead to incorrect results. For example, with the pattern
961 /1234/, partially matching against "123" and then "a4" gave a match.
963 2. Changes to pcregrep:
965 (a) All non-match returns from pcre_exec() were being treated as failures
966 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
967 error message is output. Some extra information is given for the
969 probably the only errors that are likely to be caused by users (by
970 specifying a regex that has nested indefinite repeats, for instance).
971 If there are more than 20 of these errors, pcregrep is abandoned.
973 (b) A binary zero was treated as data while matching, but terminated the
974 output line if it was written out. This has been fixed: binary zeroes
975 are now no different to any other data bytes.
977 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
978 used to set a locale for matching. The --locale=xxxx long option has
979 been added (no short equivalent) to specify a locale explicitly on the
980 pcregrep command, overriding the environment variables.
982 (d) When -B was used with -n, some line numbers in the output were one less
983 than they should have been.
985 (e) Added the -o (--only-matching) option.
987 (f) If -A or -C was used with -c (count only), some lines of context were
988 accidentally printed for the final match.
990 (g) Added the -H (--with-filename) option.
992 (h) The combination of options -rh failed to suppress file names for files
993 that were found from directory arguments.
995 (i) Added the -D (--devices) and -d (--directories) options.
997 (j) Added the -F (--fixed-strings) option.
999 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1001 (l) Added the --colo(u)r option.
1003 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1004 is not present by default.
1006 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1007 items such as (?R) or (?1), when the recursion could match a number of
1008 alternatives. If it matched one of the alternatives, but subsequently,
1009 outside the recursion, there was a failure, the code tried to back up into
1010 the recursion. However, because of the way PCRE is implemented, this is not
1011 possible, and the result was an incorrect result from the match.
1013 In order to prevent this happening, the specification of recursion has
1014 been changed so that all such subpatterns are automatically treated as
1015 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1017 4. I had overlooked the fact that, in some locales, there are characters for
1018 which isalpha() is true but neither isupper() nor islower() are true. In
1019 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1020 and ordfeminine) are like this. This affected the treatment of \w and \W
1021 when they appeared in character classes, but not when they appeared outside
1022 a character class. The bit map for "word" characters is now created
1023 separately from the results of isalnum() instead of just taking it from the
1024 upper, lower, and digit maps. (Plus the underscore character, of course.)
1026 5. The above bug also affected the handling of POSIX character classes such as
1027 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1028 permanent tables. Instead, the bit maps for such a class were previously
1029 created as the appropriate unions of the upper, lower, and digit bitmaps.
1030 Now they are created by subtraction from the [[:word:]] class, which has
1031 its own bitmap.
1033 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1034 It is created by subtracting the vertical space characters (\x09, \x0a,
1035 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1036 subtraction was done in the overall bitmap for a character class, meaning
1037 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1038 be recognized. This bug has been fixed.
1040 7. Patches from the folks at Google:
1042 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1043 real life, but is still worth protecting against".
1045 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1046 regular expressions".
1048 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1049 have it.
1051 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1052 "configure" and the latter not, in order to fix a problem somebody had
1053 with compiling the Arg class on HP-UX.
1055 (e) Improve the error-handling of the C++ wrapper a little bit.
1057 (f) New tests for checking recursion limiting.
1059 8. The pcre_memmove() function, which is used only if the environment does not
1060 have a standard memmove() function (and is therefore rarely compiled),
1061 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1062 returning a result (though PCRE never actually uses the result).
1064 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1065 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1066 returned instead of calling malloc() with an overflowing number that would
1067 most likely cause subsequent chaos.
1069 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1071 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1072 with this option is matched, the nmatch and pmatch options of regexec() are
1073 ignored.
1075 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1076 provided in case anyone wants to the the POSIX interface with UTF-8
1077 strings.
1079 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1080 C++ linking (needed for some HP-UX environments).
1082 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1083 (unused parameter) and in the pcre_printint() function (omitted "default"
1084 switch label when the default is to do nothing).
1086 15. Added some code to make it possible, when PCRE is compiled as a C++
1087 library, to replace subject pointers for pcre_exec() with a smart pointer
1088 class, thus making it possible to process discontinuous strings.
1090 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1091 much the same function. They were added by different people who were trying
1092 to make PCRE easy to compile on non-Unix systems. It has been suggested
1093 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1094 for compiling on Windows systems. I have therefore replaced it with
1095 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1096 defaults to "extern" for C or "extern C" for C++, which works fine on
1097 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1098 SCOPE with something explicit in config.h. In addition:
1100 (a) pcreposix.h still had just "extern" instead of either of these macros;
1101 I have replaced it with PCRE_DATA_SCOPE.
1103 (b) Functions such as _pcre_xclass(), which are internal to the library,
1104 but external in the C sense, all had PCRE_EXPORT in their definitions.
1105 This is apparently wrong for the Windows case, so I have removed it.
1106 (It makes no difference on Unix-like systems.)
1108 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1109 of recursive calls to match(). This is different to MATCH_LIMIT because
1110 that limits the total number of calls to match(), not all of which increase
1111 the depth of recursion. Limiting the recursion depth limits the amount of
1112 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1113 when PCRE is compiled, and changed at run time. A patch from Google adds
1114 this functionality to the C++ interface.
1116 18. Changes to the handling of Unicode character properties:
1118 (a) Updated the table to Unicode 4.1.0.
1120 (b) Recognize characters that are not in the table as "Cn" (undefined).
1122 (c) I revised the way the table is implemented to a much improved format
1123 which includes recognition of ranges. It now supports the ranges that
1124 are defined in UnicodeData.txt, and it also amalgamates other
1125 characters into ranges. This has reduced the number of entries in the
1126 table from around 16,000 to around 3,000, thus reducing its size
1127 considerably. I realized I did not need to use a tree structure after
1128 all - a binary chop search is just as efficient. Having reduced the
1129 number of entries, I extended their size from 6 bytes to 8 bytes to
1130 allow for more data.
1132 (d) Added support for Unicode script names via properties such as \p{Han}.
1134 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1135 matching that character.
1137 20. When matching a repeated Unicode property with a minimum greater than zero,
1138 (for example \pL{2,}), PCRE could look past the end of the subject if it
1139 reached it while seeking the minimum number of characters. This could
1140 happen only if some of the characters were more than one byte long, because
1141 there is a check for at least the minimum number of bytes.
1143 21. Refactored the implementation of \p and \P so as to be more general, to
1144 allow for more different types of property in future. This has changed the
1145 compiled form incompatibly. Anybody with saved compiled patterns that use
1146 \p or \P will have to recompile them.
1148 22. Added "Any" and "L&" to the supported property types.
1150 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1151 but give a compile time error if the value is greater than 0xff.
1153 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1154 accidentally not being installed or uninstalled.
1156 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1157 made were to insert the current release number. This seemed silly, because
1158 it made things harder for people building PCRE on systems that don't run
1159 "configure". I have turned pcre.h into a distributed file, no longer built
1160 by "configure", with the version identification directly included. There is
1161 no longer a pcre.h.in file.
1163 However, this change necessitated a change to the pcre-config script as
1164 well. It is built from pcre-config.in, and one of the substitutions was the
1165 release number. I have updated configure.ac so that ./configure now finds
1166 the release number by grepping pcre.h.
1168 26. Added the ability to run the tests under valgrind.
1171 Version 6.4 05-Sep-05
1172 ---------------------
1174 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1175 "--" to be printed when multiple files were scanned, even when none of the
1176 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1177 consider it to be a bug, and have restored the previous behaviour.
1179 2. A couple of code tidies to get rid of compiler warnings.
1181 3. The pcretest program used to cheat by referring to symbols in the library
1182 whose names begin with _pcre_. These are internal symbols that are not
1183 really supposed to be visible externally, and in some environments it is
1184 possible to suppress them. The cheating is now confined to including
1185 certain files from the library's source, which is a bit cleaner.
1187 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1188 file's purpose clearer.
1190 5. Reorganized pcre_ucp_findchar().
1193 Version 6.3 15-Aug-05
1194 ---------------------
1196 1. The file libpcre.pc.in did not have general read permission in the tarball.
1198 2. There were some problems when building without C++ support:
1200 (a) If C++ support was not built, "make install" and "make test" still
1201 tried to test it.
1203 (b) There were problems when the value of CXX was explicitly set. Some
1204 changes have been made to try to fix these, and ...
1206 (c) --disable-cpp can now be used to explicitly disable C++ support.
1208 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1209 backslash in a target when C++ was disabled. This confuses some
1210 versions of "make", apparently. Using an intermediate variable solves
1211 this. (Same for CPP_LOBJ.)
1213 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1214 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1215 necessary on certain architectures.
1217 4. Added a setting of -export-symbols-regex to the link command to remove
1218 those symbols that are exported in the C sense, but actually are local
1219 within the library, and not documented. Their names all begin with
1220 "_pcre_". This is not a perfect job, because (a) we have to except some
1221 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1222 available (and never for static libraries). I have made a note to try to
1223 find a way round (a) in the future.
1226 Version 6.2 01-Aug-05
1227 ---------------------
1229 1. There was no test for integer overflow of quantifier values. A construction
1230 such as {1111111111111111} would give undefined results. What is worse, if
1231 a minimum quantifier for a parenthesized subpattern overflowed and became
1232 negative, the calculation of the memory size went wrong. This could have
1233 led to memory overwriting.
1235 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1237 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1238 operating environments where this matters.
1240 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1241 PCRE options from within the C++ wrapper.
1243 5. Named capturing subpatterns were not being correctly counted when a pattern
1244 was compiled. This caused two problems: (a) If there were more than 100
1245 such subpatterns, the calculation of the memory needed for the whole
1246 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1247 back references of the form \12, where the number was greater than 9, were
1248 not recognized as back references, even though there were sufficient
1249 previous subpatterns.
1251 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1252 versions of gcc, e.g. 2.95.4.
1255 Version 6.1 21-Jun-05
1256 ---------------------
1258 1. There was one reference to the variable "posix" in pcretest.c that was not
1259 surrounded by "#if !defined NOPOSIX".
1261 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1262 the cross-check on the old pcre_info() function, for the benefit of the
1263 cut-down version of PCRE that is currently imported into Exim.
1265 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1266 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1267 patterns that start (?i) but I don't think that matters. The use of (?i) is
1268 just an example; this all applies to the other options as well.
1270 4. Since libtool seems to echo the compile commands it is issuing, the output
1271 from "make" can be reduced a bit by putting "@" in front of each libtool
1272 compile command.
1274 5. Patch from the folks at Google for configure.in to be a bit more thorough
1275 in checking for a suitable C++ installation before trying to compile the
1276 C++ stuff. This should fix a reported problem when a compiler was present,
1277 but no suitable headers.
1279 6. The man pages all had just "PCRE" as their title. I have changed them to
1280 be the relevant file name. I have also arranged that these names are
1281 retained in the file doc/pcre.txt, which is a concatenation in text format
1282 of all the man pages except the little individual ones for each function.
1284 7. The NON-UNIX-USE file had not been updated for the different set of source
1285 files that come with release 6. I also added a few comments about the C++
1286 wrapper.
1289 Version 6.0 07-Jun-05
1290 ---------------------
1292 1. Some minor internal re-organization to help with my DFA experiments.
1294 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1295 didn't matter for the library itself when fully configured, but did matter
1296 when compiling without UCP support, or within Exim, where the ucp files are
1297 not imported.
1299 3. Refactoring of the library code to split up the various functions into
1300 different source modules. The addition of the new DFA matching code (see
1301 below) to a single monolithic source would have made it really too
1302 unwieldy, quite apart from causing all the code to be include in a
1303 statically linked application, when only some functions are used. This is
1304 relevant even without the DFA addition now that patterns can be compiled in
1305 one application and matched in another.
1307 The downside of splitting up is that there have to be some external
1308 functions and data tables that are used internally in different modules of
1309 the library but which are not part of the API. These have all had their
1310 names changed to start with "_pcre_" so that they are unlikely to clash
1311 with other external names.
1313 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1314 a different (DFA) algorithm. Although it is slower than the original
1315 function, it does have some advantages for certain types of matching
1316 problem.
1318 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1319 including restarting after a partial match.
1321 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1322 defined when compiling for Windows was sent to me. I have put it into the
1323 code, though I have no means of testing or verifying it.
1325 7. Added the pcre_refcount() auxiliary function.
1327 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1328 match before or at the first newline in the subject string. In pcretest,
1329 the /f option on a pattern can be used to set this.
1331 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1332 would behave wrongly. This has been present in PCRE since release 4.0.
1334 10. A number of changes to the pcregrep command:
1336 (a) Refactored how -x works; insert ^(...)$ instead of setting
1337 PCRE_ANCHORED and checking the length, in preparation for adding
1338 something similar for -w.
1340 (b) Added the -w (match as a word) option.
1342 (c) Refactored the way lines are read and buffered so as to have more
1343 than one at a time available.
1345 (d) Implemented a pcregrep test script.
1347 (e) Added the -M (multiline match) option. This allows patterns to match
1348 over several lines of the subject. The buffering ensures that at least
1349 8K, or the rest of the document (whichever is the shorter) is available
1350 for matching (and similarly the previous 8K for lookbehind assertions).
1352 (f) Changed the --help output so that it now says
1354 -w, --word-regex(p)
1356 instead of two lines, one with "regex" and the other with "regexp"
1357 because that confused at least one person since the short forms are the
1358 same. (This required a bit of code, as the output is generated
1359 automatically from a table. It wasn't just a text change.)
1361 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1362 option but starts with a hyphen. Could be a pattern or a path name
1363 starting with a hyphen, for instance.
1365 (h) "-" can be given as a file name to represent stdin.
1367 (i) When file names are being printed, "(standard input)" is used for
1368 the standard input, for compatibility with GNU grep. Previously
1369 "<stdin>" was used.
1371 (j) The option --label=xxx can be used to supply a name to be used for
1372 stdin when file names are being printed. There is no short form.
1374 (k) Re-factored the options decoding logic because we are going to add
1375 two more options that take data. Such options can now be given in four
1376 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1378 (l) Added the -A, -B, and -C options for requesting that lines of context
1379 around matches be printed.
1381 (m) Added the -L option to print the names of files that do not contain
1382 any matching lines, that is, the complement of -l.
1384 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1385 continue to scan other files.
1387 (o) The -s option was incorrectly implemented. For compatibility with other
1388 greps, it now suppresses the error message for a non-existent or non-
1389 accessible file (but not the return code). There is a new option called
1390 -q that suppresses the output of matching lines, which was what -s was
1391 previously doing.
1393 (p) Added --include and --exclude options to specify files for inclusion
1394 and exclusion when recursing.
1396 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1397 Hopefully, it now does.
1399 12. Missing cast in pcre_study().
1401 13. Added an "uninstall" target to the makefile.
1403 14. Replaced "extern" in the function prototypes in Makefile.in with
1404 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1405 world, but is set differently for Windows.
1407 15. Added a second compiling function called pcre_compile2(). The only
1408 difference is that it has an extra argument, which is a pointer to an
1409 integer error code. When there is a compile-time failure, this is set
1410 non-zero, in addition to the error test pointer being set to point to an
1411 error message. The new argument may be NULL if no error number is required
1412 (but then you may as well call pcre_compile(), which is now just a
1413 wrapper). This facility is provided because some applications need a
1414 numeric error indication, but it has also enabled me to tidy up the way
1415 compile-time errors are handled in the POSIX wrapper.
1417 16. Added VPATH=.libs to the makefile; this should help when building with one
1418 prefix path and installing with another. (Or so I'm told by someone who
1419 knows more about this stuff than I do.)
1421 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1422 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1423 match everything, including newlines. This is not POSIX-compatible, but
1424 somebody wanted the feature. From pcretest it can be activated by using
1425 both the P and the s flags.
1427 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1429 19. libpcre.pc was being incorrectly installed as executable.
1431 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1432 it now also looks for '\r' so that it will work unmodified on Windows.
1434 21. Added Google's contributed C++ wrapper to the distribution.
1436 22. Added some untidy missing memory free() calls in pcretest, to keep
1437 Electric Fence happy when testing.
1441 Version 5.0 13-Sep-04
1442 ---------------------
1444 1. Internal change: literal characters are no longer packed up into items
1445 containing multiple characters in a single byte-string. Each character
1446 is now matched using a separate opcode. However, there may be more than one
1447 byte in the character in UTF-8 mode.
1449 2. The pcre_callout_block structure has two new fields: pattern_position and
1450 next_item_length. These contain the offset in the pattern to the next match
1451 item, and its length, respectively.
1453 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1454 insertion of callouts before each pattern item. Added the /C option to
1455 pcretest to make use of this.
1457 4. On the advice of a Windows user, the lines
1459 #if defined(_WIN32) || defined(WIN32)
1460 _setmode( _fileno( stdout ), 0x8000 );
1461 #endif /* defined(_WIN32) || defined(WIN32) */
1463 have been added to the source of pcretest. This apparently does useful
1464 magic in relation to line terminators.
1466 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1467 for the benefit of those environments where the "b" makes a difference.
1469 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1470 to know about it. I have put a hack into configure.in that adds in code
1471 to set GCC=yes if CC=icc. This seems to end up at a point in the
1472 generated configure script that is early enough to affect the setting of
1473 compiler options, which is what is needed, but I have no means of testing
1474 whether it really works. (The user who reported this had patched the
1475 generated configure script, which of course I cannot do.)
1477 LATER: After change 22 below (new libtool files), the configure script
1478 seems to know about icc (and also ecc). Therefore, I have commented out
1479 this hack in configure.in.
1481 7. Added support for pkg-config (2 patches were sent in).
1483 8. Negated POSIX character classes that used a combination of internal tables
1484 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1485 [[:^ascii]]. Typically, they would match almost any characters. The other
1486 POSIX classes were not broken in this way.
1488 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1489 to find the match, as PCRE was deluded into thinking that the match had to
1490 start at the start point or following a newline. The same bug applied to
1491 patterns with negative forward assertions or any backward assertions
1492 preceding ".*" at the start, unless the pattern required a fixed first
1493 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1495 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1496 starting at the last subject character, bytes beyond the end of the subject
1497 string were read.
1499 11. Renamed the variable "class" as "classbits" to make life easier for C++
1500 users. (Previously there was a macro definition, but it apparently wasn't
1501 enough.)
1503 12. Added the new field "tables" to the extra data so that tables can be passed
1504 in at exec time, or the internal tables can be re-selected. This allows
1505 a compiled regex to be saved and re-used at a later time by a different
1506 program that might have everything at different addresses.
1508 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1509 -R library as well as a -L library.
1511 14. The debugging options of pcretest (-d on the command line or D on a
1512 pattern) showed incorrect output for anything following an extended class
1513 that contained multibyte characters and which was followed by a quantifier.
1515 15. Added optional support for general category Unicode character properties
1516 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1517 support. It adds about 90K to the size of the library. The meanings of the
1518 inbuilt class escapes such as \d and \s have NOT been changed.
1520 16. Updated pcredemo.c to include calls to free() to release the memory for the
1521 compiled pattern.
1523 17. The generated file chartables.c was being created in the source directory
1524 instead of in the building directory. This caused the build to fail if the
1525 source directory was different from the building directory, and was
1526 read-only.
1528 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1529 file. No doubt somebody will tell me if they don't make sense... Also added
1530 Dan Mooney's comments about building on OpenVMS.
1532 19. Added support for partial matching via the PCRE_PARTIAL option for
1533 pcre_exec() and the \P data escape in pcretest.
1535 20. Extended pcretest with 3 new pattern features:
1537 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1538 write the compiled pattern to the file whose name is "rest-of-line".
1539 This is a straight binary dump of the data, with the saved pointer to
1540 the character tables forced to be NULL. The study data, if any, is
1541 written too. After writing, pcretest reads a new pattern.
1543 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1544 compiled pattern from the given file. There must not be any
1545 occurrences of "<" in the file name (pretty unlikely); if there are,
1546 pcretest will instead treat the initial "<" as a pattern delimiter.
1547 After reading in the pattern, pcretest goes on to read data lines as
1548 usual.
1550 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1551 and 16-bit fields in a compiled pattern, to simulate a pattern that
1552 was compiled on a host of opposite endianness.
1554 21. The pcre-exec() function can now cope with patterns that were compiled on
1555 hosts of opposite endianness, with this restriction:
1557 As for any compiled expression that is saved and used later, the tables
1558 pointer field cannot be preserved; the extra_data field in the arguments
1559 to pcre_exec() should be used to pass in a tables address if a value
1560 other than the default internal tables were used at compile time.
1562 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1563 now diagnosed as an error. Previously, most of the time, a negative number
1564 would have been treated as zero, but if in addition "ovector" was passed as
1565 NULL, a crash could occur.
1567 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1568 new versions from the libtool 1.5 distribution (the last one is a copy of
1569 a file called libtool.m4). This seems to have fixed the need to patch
1570 "configure" to support Darwin 1.3 (which I used to do). However, I still
1571 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1572 workstation).
1574 24. Changed the PCRE licence to be the more standard "BSD" licence.
1577 Version 4.5 01-Dec-03
1578 ---------------------
1580 1. There has been some re-arrangement of the code for the match() function so
1581 that it can be compiled in a version that does not call itself recursively.
1582 Instead, it keeps those local variables that need separate instances for
1583 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1584 needs to "recurse". Keeping track of where control must go is done by means
1585 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1586 hide most of the details from the main code, and operates only if
1587 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1588 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1589 operating.
1591 To make it easier for callers to provide specially tailored get/free
1592 functions for this usage, two new functions, pcre_stack_malloc, and
1593 pcre_stack_free, are used. They are always called in strict stacking order,
1594 and the size of block requested is always the same.
1596 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1597 PCRE has been compiled to use the stack or the heap for recursion. The
1598 -C option of pcretest uses this to show which version is compiled.
1600 A new data escape \S, is added to pcretest; it causes the amounts of store
1601 obtained and freed by both kinds of malloc/free at match time to be added
1602 to the output.
1604 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1605 what's available on my current Linux desktop machine.
1607 3. When matching a UTF-8 string, the test for a valid string at the start has
1608 been extended. If start_offset is not zero, PCRE now checks that it points
1609 to a byte that is the start of a UTF-8 character. If not, it returns
1610 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1611 this is necessary because there may be backward assertions in the pattern.
1612 When matching the same subject several times, it may save resources to use
1613 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1615 4. The code for checking the validity of UTF-8 strings has been tightened so
1616 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1617 containing "overlong sequences".
1619 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1620 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1621 should have been "&", but it just so happened that all the cases this let
1622 through by mistake were picked up later in the function.
1624 6. I had used a variable called "isblank" - this is a C99 function, causing
1625 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1627 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1628 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1629 so the version is shown; (c) stop "make test" echoing "./RunTest".
1631 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1633 9. The prototype for memmove() for systems that don't have it was using
1634 size_t, but the inclusion of the header that defines size_t was later. I've
1635 moved the #includes for the C headers earlier to avoid this.
1637 10. Added some adjustments to the code to make it easier to compiler on certain
1638 special systems:
1640 (a) Some "const" qualifiers were missing.
1641 (b) Added the macro EXPORT before all exported functions; by default this
1642 is defined to be empty.
1643 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1644 that it reads its output file name as an argument instead of writing
1645 to the standard output and assuming this can be redirected.
1647 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1648 class containing characters with values greater than 255, PCRE compilation
1649 went into a loop.
1651 12. A recursive reference to a subpattern that was within another subpattern
1652 that had a minimum quantifier of zero caused PCRE to crash. For example,
1653 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1654 recursion. If the recursively-called subpattern itself had a zero repeat,
1655 that was OK.
1657 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1658 buffer into which it was copied (for escape processing) was still set at
1659 1024, so long lines caused crashes.
1661 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1662 "internal error: code overflow...". This applied to any character class
1663 that was followed by a possessive quantifier.
1665 15. Modified the Makefile to add libpcre.la as a prerequisite for
1666 libpcreposix.la because I was told this is needed for a parallel build to
1667 work.
1669 16. If a pattern that contained .* following optional items at the start was
1670 studied, the wrong optimizing data was generated, leading to matching
1671 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1672 matching string must start with a or b or c. The correct conclusion for
1673 this pattern is that a match can start with any character.
1676 Version 4.4 13-Aug-03
1677 ---------------------
1679 1. In UTF-8 mode, a character class containing characters with values between
1680 127 and 255 was not handled correctly if the compiled pattern was studied.
1681 In fixing this, I have also improved the studying algorithm for such
1682 classes (slightly).
1684 2. Three internal functions had redundant arguments passed to them. Removal
1685 might give a very teeny performance improvement.
1687 3. Documentation bug: the value of the capture_top field in a callout is *one
1688 more than* the number of the hightest numbered captured substring.
1690 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1691 in incorrectly linking with a previously installed version. They now link
1692 explicitly with libpcre.la.
1694 5. configure.in no longer needs to recognize Cygwin specially.
1696 6. A problem in pcre.in for Windows platforms is fixed.
1698 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1699 pcretest, it used to include the size of the study block as part of its
1700 output. Unfortunately, the structure contains a field that has a different
1701 size on different hardware architectures. This meant that the tests that
1702 showed this size failed. As the block is currently always of a fixed size,
1703 this information isn't actually particularly useful in pcretest output, so
1704 I have just removed it.
1706 8. Three pre-processor statements accidentally did not start in column 1.
1707 Sadly, there are *still* compilers around that complain, even though
1708 standard C has not required this for well over a decade. Sigh.
1710 9. In pcretest, the code for checking callouts passed small integers in the
1711 callout_data field, which is a void * field. However, some picky compilers
1712 complained about the casts involved for this on 64-bit systems. Now
1713 pcretest passes the address of the small integer instead, which should get
1714 rid of the warnings.
1716 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1717 both compile and run time, and gives an error if an invalid UTF-8 sequence
1718 is found. There is a option for disabling this check in cases where the
1719 string is known to be correct and/or the maximum performance is wanted.
1721 11. In response to a bug report, I changed one line in Makefile.in from
1723 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1724 to
1725 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1727 to look similar to other lines, but I have no way of telling whether this
1728 is the right thing to do, as I do not use Windows. No doubt I'll get told
1729 if it's wrong...
1732 Version 4.3 21-May-03
1733 ---------------------
1735 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1736 Makefile.
1738 2. Some refactoring to improve the quality of the code:
1740 (i) The utf8_table... variables are now declared "const".
1742 (ii) The code for \cx, which used the "case flipping" table to upper case
1743 lower case letters, now just substracts 32. This is ASCII-specific,
1744 but the whole concept of \cx is ASCII-specific, so it seems
1745 reasonable.
1747 (iii) PCRE was using its character types table to recognize decimal and
1748 hexadecimal digits in the pattern. This is silly, because it handles
1749 only 0-9, a-f, and A-F, but the character types table is locale-
1750 specific, which means strange things might happen. A private
1751 table is now used for this - though it costs 256 bytes, a table is
1752 much faster than multiple explicit tests. Of course, the standard
1753 character types table is still used for matching digits in subject
1754 strings against \d.
1756 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1757 ending in _t are). So I've renamed it as ESC_tee.
1759 3. The first argument for regexec() in the POSIX wrapper should have been
1760 defined as "const".
1762 4. Changed pcretest to use malloc() for its buffers so that they can be
1763 Electric Fenced for debugging.
1765 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1766 to read one or more bytes before the start of the subject string. Often this
1767 had no effect on PCRE's behaviour, but in some circumstances it could
1768 provoke a segmentation fault.
1770 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1771 to try to read one or more bytes before the start of the subject string.
1773 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1774 UTF-8 support could misbehave in various ways if the subject string
1775 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1776 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1777 back over UTF-8 characters.)
1780 Version 4.2 14-Apr-03
1781 ---------------------
1783 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1785 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1786 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1787 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1788 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1789 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1791 Note: automatic setting of the BUILD variables is not yet working
1792 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1793 compile-time but not at link-time
1794 [LINK]: use for linking executables only
1795 make different versions for Windows and non-Windows
1796 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1797 libraries
1798 [LINK_FOR_BUILD]: new variable
1799 [OBJEXT]: use throughout
1800 [EXEEXT]: use throughout
1801 <winshared>: new target
1802 <wininstall>: new target
1803 <dftables.o>: use native compiler
1804 <dftables>: use native linker
1805 <install>: handle Windows platform correctly
1806 <clean>: ditto
1807 <check>: ditto
1808 copy DLL to top builddir before testing
1810 As part of these changes, -no-undefined was removed again. This was reported
1811 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1812 in any case.
1814 3. Some tidies to get rid of compiler warnings:
1816 . In the match_data structure, match_limit was an unsigned long int, whereas
1817 match_call_count was an int. I've made them both unsigned long ints.
1819 . In pcretest the fact that a const uschar * doesn't automatically cast to
1820 a void * provoked a warning.
1822 . Turning on some more compiler warnings threw up some "shadow" variables
1823 and a few more missing casts.
1825 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1826 option, a class that contained a single character with a value between 128
1827 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1829 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1830 option, a class that contained several characters, but with at least one
1831 whose value was between 128 and 255 caused PCRE to crash.
1834 Version 4.1 12-Mar-03
1835 ---------------------
1837 1. Compiling with gcc -pedantic found a couple of places where casts were
1838 needed, and a string in dftables.c that was longer than standard compilers are
1839 required to support.
1841 2. Compiling with Sun's compiler found a few more places where the code could
1842 be tidied up in order to avoid warnings.
1844 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1845 first of these names is deprecated in the latest Autoconf in favour of the name
1846 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1847 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1848 analogy I have changed it to CFLAGS_FOR_BUILD.
1850 4. Added -no-undefined to the linking command in the Makefile, because this is
1851 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1852 linking step for the pcreposix library.
1854 5. PCRE was failing to diagnose the case of two named groups with the same
1855 name.
1857 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1858 literal character that is needed in the subject for a match, and scans along to
1859 ensure that it is present before embarking on the full matching process. This
1860 saves time in cases of nested unlimited repeats that are never going to match.
1861 Problem: the scan can take a lot of time if the subject is very long (e.g.
1862 megabytes), thus penalizing straightforward matches. It is now done only if the
1863 amount of subject to be scanned is less than 1000 bytes.
1865 7. A lesser problem with the same optimization is that it was recording the
1866 first character of an anchored pattern as "needed", thus provoking a search
1867 right along the subject, even when the first match of the pattern was going to
1868 fail. The "needed" character is now not set for anchored patterns, unless it
1869 follows something in the pattern that is of non-fixed length. Thus, it still
1870 fulfils its original purpose of finding quick non-matches in cases of nested
1871 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1874 Version 4.0 17-Feb-03
1875 ---------------------
1877 1. If a comment in an extended regex that started immediately after a meta-item
1878 extended to the end of string, PCRE compiled incorrect data. This could lead to
1879 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1881 2. Moved to autoconf 2.53 and libtool 1.4.2.
1883 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1884 the special perltest8 script is no longer needed - all the tests can be run
1885 from a single perltest script.
1887 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1888 by \s. It has now been removed in PCRE. This means it isn't recognized as
1889 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1890 class [:space:] *does* include VT, thereby creating a mess.
1892 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1893 space and tab.
1895 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1896 its new features into the main test script, reducing the number of scripts.
1898 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1899 were backward compatible, and made the (?i) apply to the whole pattern, as if
1900 /i were given. Now it behaves more logically, and applies the option setting
1901 only to what follows. PCRE has been changed to follow suit. However, if it
1902 finds options settings right at the start of the pattern, it extracts them into
1903 the global options, as before. Thus, they show up in the info data.
1905 8. Added support for the \Q...\E escape sequence. Characters in between are
1906 treated as literals. This is slightly different from Perl in that $ and @ are
1907 also handled as literals inside the quotes. In Perl, they will cause variable
1908 interpolation. Note the following examples:
1910 Pattern PCRE matches Perl matches
1912 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1913 \Qabc\$xyz\E abc\$xyz abc\$xyz
1914 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1916 For compatibility with Perl, \Q...\E sequences are recognized inside character
1917 classes as well as outside them.
1919 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1920 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1921 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1922 signed/unsigned warnings.
1924 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1925 option for pcretest, so I've replaced it by a simple function that does just
1926 that job.
1928 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1929 "pcregrep -".
1931 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1932 Java package. This provides some syntactic sugar for simple cases of what my
1933 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1934 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1935 item, you can use this simplified notation. Note that only makes sense with
1936 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1937 greediness, whatever the setting of the PCRE_UNGREEDY option.
1939 13. A change of greediness default within a pattern was not taking effect at
1940 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1941 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1942 was abstracted outside.
1944 14. PCRE now supports the \G assertion. It is true when the current matching
1945 position is at the start point of the match. This differs from \A when the
1946 starting offset is non-zero. Used with the /g option of pcretest (or similar
1947 code), it works in the same way as it does for Perl's /g option. If all
1948 alternatives of a regex begin with \G, the expression is anchored to the start
1949 match position, and the "anchored" flag is set in the compiled expression.
1951 15. Some bugs concerning the handling of certain option changes within patterns
1952 have been fixed. These applied to options other than (?ims). For example,
1953 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1954 been the other way round. Some of this was related to change 7 above.
1956 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1957 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1958 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1959 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1961 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1962 mode. Unlike ".", it always matches newline, whatever the setting of
1963 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1964 assertions. Perl allows it, but it doesn't (in general) work because it can't
1965 calculate the length of the lookbehind. At least, that's the case for Perl
1966 5.8.0 - I've been told they are going to document that it doesn't work in
1967 future.
1969 18. Added an error diagnosis for escapes that PCRE does not support: these are
1970 \L, \l, \N, \P, \p, \U, \u, and \X.
1972 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1973 reading past the end of the pattern in cases such as /[abcd/.
1975 20. PCRE was getting more memory than necessary for patterns with classes that
1976 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1978 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1979 compiling PCRE for use with Virtual Pascal.
1981 22. Small fix to the Makefile to make it work properly if the build is done
1982 outside the source tree.
1984 23. Added a new extension: a condition to go with recursion. If a conditional
1985 subpattern starts with (?(R) the "true" branch is used if recursion has
1986 happened, whereas the "false" branch is used only at the top level.
1988 24. When there was a very long string of literal characters (over 255 bytes
1989 without UTF support, over 250 bytes with UTF support), the computation of how
1990 much memory was required could be incorrect, leading to segfaults or other
1991 strange effects.
1993 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1994 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1995 there was a subsequent back reference to those brackets. This meant that, for
1996 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1997 possible to check for precisely this case. All we can do is abandon the
1998 optimization if .* occurs inside capturing brackets when there are any back
1999 references whatsoever. (See below for a better fix that came later.)
2001 26. The handling of the optimization for finding the first character of a
2002 non-anchored pattern, and for finding a character that is required later in the
2003 match were failing in some cases. This didn't break the matching; it just
2004 failed to optimize when it could. The way this is done has been re-implemented.
2006 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2008 28. Added a new feature that provides some of the functionality that Perl
2009 provides with (?{...}). The facility is termed a "callout". The way it is done
2010 in PCRE is for the caller to provide an optional function, by setting
2011 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2012 global variable. By default it is unset, which disables all calling out. To get
2013 the function called, the regex must include (?C) at appropriate points. This
2014 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2015 This provides a means of identifying different callout points. When PCRE
2016 reaches such a point in the regex, if pcre_callout has been set, the external
2017 function is called. It is provided with data in a structure called
2018 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2019 matching continues; if it returns a non-zero value, the match at the current
2020 point fails. However, backtracking will occur if possible. [This was changed
2021 later and other features added - see item 49 below.]
2023 29. pcretest is upgraded to test the callout functionality. It provides a
2024 callout function that displays information. By default, it shows the start of
2025 the match and the current position in the text. There are some new data escapes
2026 to vary what happens:
2028 \C+ in addition, show current contents of captured substrings
2029 \C- do not supply a callout function
2030 \C!n return 1 when callout number n is reached
2031 \C!n!m return 1 when callout number n is reached for the mth time
2033 30. If pcregrep was called with the -l option and just a single file name, it
2034 output "<stdin>" if a match was found, instead of the file name.
2036 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2037 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2038 pcre_exec(). This saves a malloc/free per call. The default value of
2039 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2040 when configuring.
2042 32. The default maximum size of a compiled pattern is 64K. There have been a
2043 few cases of people hitting this limit. The code now uses macros to handle the
2044 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2045 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2046 configuring. Tests 2 and 5 work only with 2-byte links because they output
2047 debugging information about compiled patterns.
2049 33. Internal code re-arrangements:
2051 (a) Moved the debugging function for printing out a compiled regex into
2052 its own source file (printint.c) and used #include to pull it into
2053 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2054 separate copies.
2056 (b) Defined the list of op-code names for debugging as a macro in
2057 internal.h so that it is next to the definition of the opcodes.
2059 (c) Defined a table of op-code lengths for simpler skipping along compiled
2060 code. This is again a macro in internal.h so that it is next to the
2061 definition of the opcodes.
2063 34. Added support for recursive calls to individual subpatterns, along the
2064 lines of Robin Houston's patch (but implemented somewhat differently).
2066 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2067 allow it to read and process whole directories in Win32. This code was
2068 contributed by Lionel Fourquaux; it has not been tested by me.
2070 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2071 used to name a group. Names consist of alphanumerics and underscores, and must
2072 be unique. Back references use the syntax (?P=name) and recursive calls use
2073 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2074 numbers. The function pcre_fullinfo() can be used after compilation to extract
2075 a name/number map. There are three relevant calls:
2077 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2078 PCRE_INFO_NAMECOUNT yields the number of entries
2079 PCRE_INFO_NAMETABLE yields a pointer to the map.
2081 The map is a vector of fixed-size entries. The size of each entry depends on
2082 the length of the longest name used. The first two bytes of each entry are the
2083 group number, most significant byte first. There follows the corresponding
2084 name, zero terminated. The names are in alphabetical order.
2086 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2087 case instead of 255. Making it the same both with and without UTF-8 support
2088 means that the same test output works with both.
2090 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2091 calling malloc() with a zero argument.
2093 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2094 optimization. I've improved things by keeping a bitmap of backreferences with
2095 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2096 fact referenced, the optimization can be applied. It is unlikely that a
2097 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2098 the match to follow \n) will appear inside brackets with a number greater than
2099 31, but if it does, any back reference > 31 suppresses the optimization.
2101 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2102 of disabling numbered capturing parentheses. Any opening parenthesis that is
2103 not followed by ? behaves as if it were followed by ?: but named parentheses
2104 can still be used for capturing (and they will acquire numbers in the usual
2105 way).
2107 41. Redesigned the return codes from the match() function into yes/no/error so
2108 that errors can be passed back from deep inside the nested calls. A malloc
2109 failure while inside a recursive subpattern call now causes the
2110 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2112 42. It is now possible to set a limit on the number of times the match()
2113 function is called in a call to pcre_exec(). This facility makes it possible to
2114 limit the amount of recursion and backtracking, though not in a directly
2115 obvious way, because the match() function is used in a number of different
2116 circumstances. The count starts from zero for each position in the subject
2117 string (for non-anchored patterns). The default limit is, for compatibility, a
2118 large number, namely 10 000 000. You can change this in two ways:
2120 (a) When configuring PCRE before making, you can use --with-match-limit=n
2121 to set a default value for the compiled library.
2123 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2124 a different value is set. See 45 below.
2126 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2128 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2129 of things that can be changed at compile time. The first argument specifies
2130 what is wanted and the second points to where the information is to be placed.
2131 The current list of available information is:
2135 The output is an integer that is set to one if UTF-8 support is available;
2136 otherwise it is set to zero.
2140 The output is an integer that it set to the value of the code that is used for
2141 newline. It is either LF (10) or CR (13).
2145 The output is an integer that contains the number of bytes used for internal
2146 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2150 The output is an integer that contains the threshold above which the POSIX
2151 interface uses malloc() for output vectors. See item 31 above.
2155 The output is an unsigned integer that contains the default limit of the number
2156 of match() calls in a pcre_exec() execution. See 42 above.
2158 44. pcretest has been upgraded by the addition of the -C option. This causes it
2159 to extract all the available output from the new pcre_config() function, and to
2160 output it. The program then exits immediately.
2162 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2163 order to support additional features. One way would have been to define
2164 pcre_exec2() (for example) with extra arguments, but this would not have been
2165 extensible, and would also have required all calls to the original function to
2166 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2167 is used for passing in "extra" data from pcre_study().
2169 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2170 contains the following fields:
2172 flags a bitmap indicating which of the following fields are set
2173 study_data opaque data from pcre_study()
2174 match_limit a way of specifying a limit on match() calls for a specific
2175 call to pcre_exec()
2176 callout_data data for callouts (see 49 below)
2178 The flag bits are also defined in pcre.h, and are
2184 The pcre_study() function now returns one of these new pcre_extra blocks, with
2185 the actual study data pointed to by the study_data field, and the
2186 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2187 before. That is, this change is entirely upwards-compatible and requires no
2188 change to existing code.
2190 If you want to pass in additional data to pcre_exec(), you can either place it
2191 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2192 block.
2194 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2195 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2196 times with different match limits, until it finds the minimum value needed for
2197 pcre_exec() to complete. The value is then output. This can be instructive; for
2198 most simple matches the number is quite small, but for pathological cases it
2199 gets very large very quickly.
2201 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2202 returns the size of the data block pointed to by the study_data field in a
2203 pcre_extra block, that is, the value that was passed as the argument to
2204 pcre_malloc() when PCRE was getting memory in which to place the information
2205 created by pcre_study(). The fourth argument should point to a size_t variable.
2206 pcretest has been extended so that this information is shown after a successful
2207 pcre_study() call when information about the compiled regex is being displayed.
2209 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2210 because what follows is always an absolute path. (Later: it turns out that this
2211 is more than cosmetic for MinGW, because it doesn't like empty path
2212 components.)
2214 49. Some changes have been made to the callout feature (see 28 above):
2216 (i) A callout function now has three choices for what it returns:
2218 0 => success, carry on matching
2219 > 0 => failure at this point, but backtrack if possible
2220 < 0 => serious error, return this value from pcre_exec()
2222 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2223 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2224 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2225 use by callout functions. It will never be used by PCRE itself.
2227 (ii) The pcre_extra structure (see 45 above) has a void * field called
2228 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2229 pcre_callout_block structure has a field of the same name. The contents of
2230 the field passed in the pcre_extra structure are passed to the callout
2231 function in the corresponding field in the callout block. This makes it
2232 easier to use the same callout-containing regex from multiple threads. For
2233 testing, the pcretest program has a new data escape
2235 \C*n pass the number n (may be negative) as callout_data
2237 If the callout function in pcretest receives a non-zero value as
2238 callout_data, it returns that value.
2240 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2241 there were some redundant $(CFLAGS) in commands that are now specified as
2242 $(LINK), which already includes $(CFLAGS).
2244 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2245 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2246 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2247 one-byte characters throughout. Note that case-insensitive matching applies
2248 only to characters whose values are less than 256. PCRE doesn't support the
2249 notion of cases for higher-valued characters.
2251 (i) A character class whose characters are all within 0-255 is handled as
2252 a bit map, and the map is inverted for negative classes. Previously, a
2253 character > 255 always failed to match such a class; however it should
2254 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2256 (ii) A negated character class with a single character < 255 is coded as
2257 "not this character" (OP_NOT). This wasn't working properly when the test
2258 character was multibyte, either singly or repeated.
2260 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2261 mode, for example: \x{100}{2,3}.
2263 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2264 singly or repeated) now correctly test multibyte characters. However,
2265 PCRE doesn't recognize any characters with values greater than 255 as
2266 digits, spaces, or word characters. Such characters always match \D, \S,
2267 and \W, and never match \d, \s, or \w.
2269 (v) Classes may now contain characters and character ranges with values
2270 greater than 255. For example: [ab\x{100}-\x{400}].
2272 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2273 PCRE in UTF-8 mode.
2275 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2276 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2277 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2278 value.)
2280 53. The single man page has become too large. I have therefore split it up into
2281 a number of separate man pages. These also give rise to individual HTML pages;
2282 these are now put in a separate directory, and there is an index.html page that
2283 lists them all. Some hyperlinking between the pages has been installed.
2285 54. Added convenience functions for handling named capturing parentheses.
2287 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2288 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2289 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2290 are faulted.
2292 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2293 calling configure. These values are used when compiling the dftables.c program
2294 which is run to generate the source of the default character tables. They
2295 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2296 you will need to set these values.
2298 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2301 Version 3.9 02-Jan-02
2302 ---------------------
2304 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2306 2. If --disable-static was given, the building process failed when trying to
2307 build pcretest and pcregrep. (For some reason it was using libtool to compile
2308 them, which is not right, as they aren't part of the library.)
2311 Version 3.8 18-Dec-01
2312 ---------------------
2314 1. The experimental UTF-8 code was completely screwed up. It was packing the
2315 bytes in the wrong order. How dumb can you get?
2318 Version 3.7 29-Oct-01
2319 ---------------------
2321 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2322 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2323 this didn't happen under Solaris 8, where I normally test things.
2325 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2326 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2327 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2328 this fix an hour or so after the initial 3.7 release.)
2331 Version 3.6 23-Oct-01
2332 ---------------------
2334 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2335 offsets passed as NULL with zero offset count.
2337 2. The config.guess and config.sub files had not been updated when I moved to
2338 the latest autoconf.
2341 Version 3.5 15-Aug-01
2342 ---------------------
2344 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2345 had been forgotten.
2347 2. By using declared but undefined structures, we can avoid using "void"
2348 definitions in pcre.h while keeping the internal definitions of the structures
2349 private.
2351 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2352 user point of view, this means that both static and shared libraries are built
2353 by default, but this can be individually controlled. More of the work of
2354 handling this static/shared cases is now inside libtool instead of PCRE's make
2355 file.
2357 4. The pcretest utility is now installed along with pcregrep because it is
2358 useful for users (to test regexs) and by doing this, it automatically gets
2359 relinked by libtool. The documentation has been turned into a man page, so
2360 there are now .1, .txt, and .html versions in /doc.
2362 5. Upgrades to pcregrep:
2363 (i) Added long-form option names like gnu grep.
2364 (ii) Added --help to list all options with an explanatory phrase.
2365 (iii) Added -r, --recursive to recurse into sub-directories.
2366 (iv) Added -f, --file to read patterns from a file.
2368 6. pcre_exec() was referring to its "code" argument before testing that
2369 argument for NULL (and giving an error if it was NULL).
2371 7. Upgraded Makefile.in to allow for compiling in a different directory from
2372 the source directory.
2374 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2375 options bits, the pointer it was passed was to an int instead of to an unsigned
2376 long int. This mattered only on 64-bit systems.
2378 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2379 generated) instead of pcre.in, which it its source. Also made the same change
2380 in several of the .c files.
2382 10. A new release of gcc defines printf() as a macro, which broke pcretest
2383 because it had an ifdef in the middle of a string argument for printf(). Fixed
2384 by using separate calls to printf().
2386 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2387 script, to force use of CR or LF instead of \n in the source. On non-Unix
2388 systems, the value can be set in config.h.
2390 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2391 absolute limit. Changed the text of the error message to make this clear, and
2392 likewise updated the man page.
2394 13. The limit of 99 on the number of capturing subpatterns has been removed.
2395 The new limit is 65535, which I hope will not be a "real" limit.
2398 Version 3.4 22-Aug-00
2399 ---------------------
2401 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2403 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2406 Version 3.3 01-Aug-00
2407 ---------------------
2409 1. If an octal character was given, but the value was greater than \377, it
2410 was not getting masked to the least significant bits, as documented. This could
2411 lead to crashes in some systems.
2413 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2414 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2416 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2417 These just pass their arguments on to (pcre_free)(), but they are provided
2418 because some uses of PCRE bind it to non-C systems that can call its functions,
2419 but cannot call free() or pcre_free() directly.
2421 4. Add "make test" as a synonym for "make check". Corrected some comments in
2422 the Makefile.
2424 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2425 Makefile.
2427 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2428 command called pgrep for grepping around the active processes.
2430 7. Added the beginnings of support for UTF-8 character strings.
2432 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2433 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2434 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2435 out for the ar command.)
2438 Version 3.2 12-May-00
2439 ---------------------
2441 This is purely a bug fixing release.
2443 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2444 of ZA. This was just one example of several cases that could provoke this bug,
2445 which was introduced by change 9 of version 2.00. The code for breaking
2446 infinite loops after an iteration that matches an empty string was't working
2447 correctly.
2449 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2450 when matched against abbab (for example). After matching an empty string, it
2451 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2452 caused it to match further down the string than it should.
2454 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2455 was there because it doesn't seem to be needed, and it causes trouble on some
2456 systems, as it is not a Standard C header. It has been removed.
2458 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2459 were reported on the Macintosh. The changes were from
2461 while ((c = *(++ptr)) != 0 && c != '\n');
2462 to
2463 while ((c = *(++ptr)) != 0 && c != '\n') ;
2465 Totally extraordinary, but if that's what it takes...
2467 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2468 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2469 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2470 assumes the way PCRE uses memmove() (always moving upwards).
2472 6. PCRE is being used in one environment where strchr() is not available. There
2473 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2474 faster code anyway.
2477 Version 3.1 09-Feb-00
2478 ---------------------
2480 The only change in this release is the fixing of some bugs in Makefile.in for
2481 the "install" target:
2483 (1) It was failing to install pcreposix.h.
2485 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2488 Version 3.0 01-Feb-00
2489 ---------------------
2491 1. Add support for the /+ modifier to perltest (to output $` like it does in
2492 pcretest).
2494 2. Add support for the /g modifier to perltest.
2496 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2497 matches null strings.
2499 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2500 pattern. Perl treats empty patterns specially - it reuses the most recent
2501 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2502 effect.
2504 5. The POSIX interface was broken in that it was just handing over the POSIX
2505 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2506 required a bigger vector, with some working space on the end. This means that
2507 the POSIX wrapper now has to get and free some memory, and copy the results.
2509 6. Added some simple autoconf support, placing the test data and the
2510 documentation in separate directories, re-organizing some of the
2511 information files, and making it build pcre-config (a GNU standard). Also added
2512 libtool support for building PCRE as a shared library, which is now the
2513 default.
2515 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2516 09 are not valid octal constants. Single digits will be used for minor values
2517 less than 10.
2519 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2520 existing programs that set these in the POSIX interface can use PCRE without
2521 modification.
2523 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2524 return all that pcre_info() returns, plus additional data. The pcre_info()
2525 function is retained for compatibility, but is considered to be obsolete.
2527 10. Added experimental recursion feature (?R) to handle one common case that
2528 Perl 5.6 will be able to do with (?p{...}).
2530 11. Added support for POSIX character classes like [:alpha:], which Perl is
2531 adopting.
2534 Version 2.08 31-Aug-99
2535 ----------------------
2537 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2538 trying to match at the startoffset position, but instead was moving forward to
2539 the next newline as if a previous match had failed.
2541 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2542 and could get into a loop if a null string was matched other than at the start
2543 of the subject.
2545 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2546 be distinguished at compile time, and for completeness also added PCRE_DATE.
2548 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2549 in GnuWin32 environments.
2552 Version 2.07 29-Jul-99
2553 ----------------------
2555 1. The documentation is now supplied in plain text form and HTML as well as in
2556 the form of man page sources.
2558 2. C++ compilers don't like assigning (void *) values to other pointer types.
2559 In particular this affects malloc(). Although there is no problem in Standard
2560 C, I've put in casts to keep C++ compilers happy.
2562 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2563 should be (const char *).
2565 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2566 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2567 However, I haven't made this a standard facility. The documentation doesn't
2568 mention it, and the Makefile doesn't support it.
2570 5. The Makefile now contains an "install" target, with editable destinations at
2571 the top of the file. The pcretest program is not installed.
2573 6. pgrep -V now gives the PCRE version number and date.
2575 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2576 causing the entire string to be ignored, instead of just the last character.
2578 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2579 non-matching string, it can take a very, very long time, even for strings of
2580 quite modest length, because of the nested recursion. PCRE now does better in
2581 some of these cases. It does this by remembering the last required literal
2582 character in the pattern, and pre-searching the subject to ensure it is present
2583 before running the real match. In other words, it applies a heuristic to detect
2584 some types of certain failure quickly, and in the above example, if presented
2585 with a string that has no trailing " it gives "no match" very quickly.
2587 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2588 other alternatives are tried instead.
2591 Version 2.06 09-Jun-99
2592 ----------------------
2594 1. Change pcretest's output for amount of store used to show just the code
2595 space, because the remainder (the data block) varies in size between 32-bit and
2596 64-bit systems.
2598 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2599 start matching at. This allows lookbehinds to work when searching for multiple
2600 occurrences in a string.
2602 3. Added additional options to pcretest for testing multiple occurrences:
2604 /+ outputs the rest of the string that follows a match
2605 /g loops for multiple occurrences, using the new startoffset argument
2606 /G loops for multiple occurrences by passing an incremented pointer
2608 4. PCRE wasn't doing the "first character" optimization for patterns starting
2609 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2610 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2611 the letter 'x'. On long subject strings, this gives a significant speed-up.
2614 Version 2.05 21-Apr-99
2615 ----------------------
2617 1. Changed the type of magic_number from int to long int so that it works
2618 properly on 16-bit systems.
2620 2. Fixed a bug which caused patterns starting with .* not to work correctly
2621 when the subject string contained newline characters. PCRE was assuming
2622 anchoring for such patterns in all cases, which is not correct because .* will
2623 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2624 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2625 must be retried after every newline in the subject.
2628 Version 2.04 18-Feb-99
2629 ----------------------
2631 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2632 computation of the store needed to hold the pattern was incorrect (too large).
2633 If such patterns were nested a few deep, this could multiply and become a real
2634 problem.
2636 2. Added /M option to pcretest to show the memory requirement of a specific
2637 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2639 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2640 compiled in such a way that the backtracking after subsequent failure was
2641 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2642 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2645 Version 2.03 02-Feb-99
2646 ----------------------
2648 1. Fixed typo and small mistake in man page.
2650 2. Added 4th condition (GPL supersedes if conflict) and created separate
2651 LICENCE file containing the conditions.
2653 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2654 Perl, that is the internal \ allows the delimiter to be included in the
2655 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2656 the final delimiter, add \ to the end of the pattern (to test the error).
2658 4. Added the convenience functions for extracting substrings after a successful
2659 match. Updated pcretest to make it able to test these functions.
2662 Version 2.02 14-Jan-99
2663 ----------------------
2665 1. Initialized the working variables associated with each extraction so that
2666 their saving and restoring doesn't refer to uninitialized store.
2668 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2669 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2670 fix the problem.
2672 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2673 calls, and wasn't printing the correct value for compiling calls. Increased the
2674 default value of LOOPREPEAT, and the number of significant figures in the
2675 times.
2677 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2679 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2680 a building problem on Windows NT with a FAT file system.
2683 Version 2.01 21-Oct-98
2684 ----------------------
2686 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2687 to character tables built by pcre_maketables() in the current locale. If NULL
2688 is passed, the default tables are used.
2691 Version 2.00 24-Sep-98
2692 ----------------------
2694 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2695 it any more.
2697 2. Allow quantification of (?>) groups, and make it work correctly.
2699 3. The first character computation wasn't working for (?>) groups.
2701 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2702 end of the subject) and add 5.005's \z, which really does match only at the
2703 very end of the subject.
2705 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2707 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2708 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2709 localized options. All options to pcre_study() were also removed.
2711 7. Add other new features from 5.005:
2713 $(?<= positive lookbehind
2714 $(?<! negative lookbehind
2715 (?imsx-imsx) added the unsetting capability
2716 such a setting is global if at outer level; local otherwise
2717 (?imsx-imsx:) non-capturing groups with option setting
2718 (?(cond)re|re) conditional pattern matching
2720 A backreference to itself in a repeated group matches the previous
2721 captured string.
2723 8. General tidying up of studying (both automatic and via "study")
2724 consequential on the addition of new assertions.
2726 9. As in 5.005, unlimited repeated groups that could match an empty substring
2727 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2728 runtime if any iteration does actually match an empty substring.
2730 10. Include the RunTest script in the distribution.
2732 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2733 discrepancies, some of which were old and were also with respect to 5.004. They
2734 have now been fixed.
2737 Version 1.09 28-Apr-98
2738 ----------------------
2740 1. A negated single character class followed by a quantifier with a minimum
2741 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2742 program crashes, or just wrong answers. This did not apply to negated classes
2743 containing more than one character, or to minima other than one.
2746 Version 1.08 27-Mar-98
2747 ----------------------
2749 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2751 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2752 latter must appear before anything that relies on it in the pattern.
2755 Version 1.07 16-Feb-98
2756 ----------------------
2758 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2759 repeat of a potentially empty string).
2762 Version 1.06 23-Jan-98
2763 ----------------------
2765 1. Added Markus Oberhumer's little patches for C++.
2767 2. Literal strings longer than 255 characters were broken.
2770 Version 1.05 23-Dec-97
2771 ----------------------
2773 1. Negated character classes containing more than one character were failing if
2774 PCRE_CASELESS was set at run time.
2777 Version 1.04 19-Dec-97
2778 ----------------------
2780 1. Corrected the man page, where some "const" qualifiers had been omitted.
2782 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2783 input syntax.
2785 3. Fixed memory leak which occurred when a regex with back references was
2786 matched with an offsets vector that wasn't big enough. The temporary memory
2787 that is used in this case wasn't being freed if the match failed.
2789 4. Tidied pcretest to ensure it frees memory that it gets.
2791 5. Temporary memory was being obtained in the case where the passed offsets
2792 vector was exactly big enough.
2794 6. Corrected definition of offsetof() from change 5 below.
2796 7. I had screwed up change 6 below and broken the rules for the use of
2797 setjmp(). Now fixed.
2800 Version 1.03 18-Dec-97
2801 ----------------------
2803 1. A erroneous regex with a missing opening parenthesis was correctly
2804 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2805 on some systems.
2807 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2808 it was reported that one broken compiler failed on the former because "code" is
2809 also an independent variable.
2811 3. The erroneous regex a[]b caused an array overrun reference.
2813 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2814 fail on data ending with that character. (It was going on too far, and checking
2815 the next character, typically a binary zero.) This was specific to the
2816 optimized code for single-character negative classes.
2818 5. Added a contributed patch from the TIN world which does the following:
2820 + Add an undef for memmove, in case the the system defines a macro for it.
2822 + Add a definition of offsetof(), in case there isn't one. (I don't know
2823 the reason behind this - offsetof() is part of the ANSI standard - but
2824 it does no harm).
2826 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2827 most of the places where whitespace preceded '#'. I have given up and
2828 allowed the remaining 2 cases to be at the margin.
2830 + Rename some variables in pcre to eliminate shadowing. This seems very
2831 pedantic, but does no harm, of course.
2833 6. Moved the call to setjmp() into its own function, to get rid of warnings
2834 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2836 7. Constructs such as \d{8,} were compiling into the equivalent of
2837 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2838 outcome, but in this particular case used more store than had been allocated,
2839 which caused the bug to be discovered because it threw up an internal error.
2841 8. The debugging code in both pcre and pcretest for outputting the compiled
2842 form of a regex was going wrong in the case of back references followed by
2843 curly-bracketed repeats.
2846 Version 1.02 12-Dec-97
2847 ----------------------
2849 1. Typos in pcre.3 and comments in the source fixed.
2851 2. Applied a contributed patch to get rid of places where it used to remove
2852 'const' from variables, and fixed some signed/unsigned and uninitialized
2853 variable warnings.
2855 3. Added the "runtest" target to Makefile.
2857 4. Set default compiler flag to -O2 rather than just -O.
2860 Version 1.01 19-Nov-97
2861 ----------------------
2863 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2864 like /([ab]*)*/, that is, for classes with more than one character in them.
2866 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2867 as /((?>a*))*/ (a PCRE_EXTRA facility).
2870 Version 1.00 18-Nov-97
2871 ----------------------
2873 1. Added compile-time macros to support systems such as SunOS4 which don't have
2874 memmove() or strerror() but have other things that can be used instead.
2876 2. Arranged that "make clean" removes the executables.
2879 Version 0.99 27-Oct-97
2880 ----------------------
2882 1. Fixed bug in code for optimizing classes with only one character. It was
2883 initializing a 32-byte map regardless, which could cause it to run off the end
2884 of the memory it had got.
2886 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2889 Version 0.98 22-Oct-97
2890 ----------------------
2892 1. Fixed bug in code for handling temporary memory usage when there are more
2893 back references than supplied space in the ovector. This could cause segfaults.
2896 Version 0.97 21-Oct-97
2897 ----------------------
2899 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2901 2. Optimized negated single characters not to use a bit map.
2903 3. Brought error texts together as macro definitions; clarified some of them;
2904 fixed one that was wrong - it said "range out of order" when it meant "invalid
2905 escape sequence".
2907 4. Changed some char * arguments to const char *.
2909 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2911 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2912 pcretest.
2915 Version 0.96 16-Oct-97
2916 ----------------------
2918 1. Added a simple "pgrep" utility to the distribution.
2920 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2921 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2922 where "ddd" means "one or more decimal digits".
2924 3. Fixed serious bug. If a pattern had a back reference, but the call to
2925 pcre_exec() didn't supply a large enough ovector to record the related
2926 identifying subpattern, the match always failed. PCRE now remembers the number
2927 of the largest back reference, and gets some temporary memory in which to save
2928 the offsets during matching if necessary, in order to ensure that
2929 backreferences always work.
2931 4. Increased the compatibility with Perl in a number of ways:
2933 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2934 to request this handling. The option can be set at compile or exec time.
2936 (b) $ matches before a terminating newline by default; an option
2937 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2938 mode). The option can be set at compile or exec time.
2940 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2941 the same as Perl's. If the decimal number it represents is less than 10
2942 or there aren't that many previous left capturing parentheses, an octal
2943 escape is read. Inside a character class, it's always an octal escape,
2944 even if it is a single digit.
2946 (d) An escaped but undefined alphabetic character is taken as a literal,
2947 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2948 escapes.
2950 (e) {0} is now permitted. (The previous item is removed from the compiled
2951 pattern).
2953 5. Changed all the names of code files so that the basic parts are no longer
2954 than 10 characters, and abolished the teeny "globals.c" file.
2956 6. Changed the handling of character classes; they are now done with a 32-byte
2957 bit map always.
2959 7. Added the -d and /D options to pcretest to make it possible to look at the
2960 internals of compilation without having to recompile pcre.
2963 Version 0.95 23-Sep-97
2964 ----------------------
2966 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2967 \x20 at the start of a run of normal characters. These were being treated as
2968 real characters, instead of the source characters being re-checked.
2971 Version 0.94 18-Sep-97
2972 ----------------------
2974 1. The functions are now thread-safe, with the caveat that the global variables
2975 containing pointers to malloc() and free() or alternative functions are the
2976 same for all threads.
2978 2. Get pcre_study() to generate a bitmap of initial characters for non-
2979 anchored patterns when this is possible, and use it if passed to pcre_exec().
2982 Version 0.93 15-Sep-97
2983 ----------------------
2985 1. /(b)|(:+)/ was computing an incorrect first character.
2987 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2988 but not actually doing anything yet.
2990 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2991 as Perl does (e.g. [-az] or [az-]).
2993 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2994 all possible positions.
2996 5. Split up into different modules to avoid including unneeded functions in a
2997 compiled binary. However, compile and exec are still in one module. The "study"
2998 function is split off.
3000 6. The character tables are now in a separate module whose source is generated
3001 by an auxiliary program - but can then be edited by hand if required. There are
3002 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3003 toupper() in the code.
3005 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3006 make them global. Abolish the function for setting them, as the caller can now
3007 set them directly.
3010 Version 0.92 11-Sep-97
3011 ----------------------
3013 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3014 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3016 2. Caseless matching was not working in character classes if the characters in
3017 the pattern were in upper case.
3019 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3021 4. Make PCRE_ANCHORED public and accept as a compile option.
3023 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3024 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3025 pass them.
3027 6. Give an error if bad option bits passed at compile or run time.
3029 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3030 pcretest to cause it to pass that flag.
3032 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3033 options, and the first character, if set.
3035 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3038 Version 0.91 10-Sep-97
3039 ----------------------
3041 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3042 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3044 2. PCRE was looping on encountering an indefinitely repeated back reference to
3045 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3046 Perl does - treats the match as successful.
3048 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5