/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 304 - (show annotations) (download)
Sun Jan 20 19:52:18 2008 UTC (6 years, 9 months ago) by ph10
File size: 149384 byte(s)
Add detail to cmake update note.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.6 19-Jan-08
5 ---------------------
6
7 1. A character class containing a very large number of characters with
8 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
9 overflow.
10
11 2. Patch to cut out the "long long" test in pcrecpp_unittest when
12 HAVE_LONG_LONG is not defined.
13
14 3. Applied Christian Ehrlicher's patch to update the CMake build files to
15 bring them up to date and include new features. This patch includes:
16
17 - Fixed PH's badly added libz and libbz2 support.
18 - Fixed a problem with static linking.
19 - Added pcredemo.
20 - Fixed dftables problem and added an option.
21 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
22 HAVE_LONG_LONG.
23 - Added readline support for pcretest.
24 - Added an listing of the option settings after cmake has run.
25
26
27 Version 7.5 10-Jan-08
28 ---------------------
29
30 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
31 values in parens when parsing an RE using the C++ wrapper."
32
33 2. Negative specials like \S did not work in character classes in UTF-8 mode.
34 Characters greater than 255 were excluded from the class instead of being
35 included.
36
37 3. The same bug as (2) above applied to negated POSIX classes such as
38 [:^space:].
39
40 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
41 defined or documented. It seems to have been a typo for PCRE_STATIC, so
42 I have changed it.
43
44 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
45 first named subpattern) and a construct such as (?&a) would reference the
46 first named subpattern whose name started with "a" (in other words, the
47 length check was missing). Both these problems are fixed. "Subpattern name
48 expected" is now given for (?&) (a zero-length name), and this patch also
49 makes it give the same error for \k'' (previously it complained that that
50 was a reference to a non-existent subpattern).
51
52 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
53 this is right because (?- can be followed by option settings as well as by
54 digits. I have, however, made the messages clearer.
55
56 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
57 than the number used in the conditional) now cause a compile-time error.
58 This is actually not compatible with Perl, which accepts such patterns, but
59 treats the conditional as always being FALSE (as PCRE used to), but it
60 seems to me that giving a diagnostic is better.
61
62 8. Change "alphameric" to the more common word "alphanumeric" in comments
63 and messages.
64
65 9. Fix two occurrences of "backslash" in comments that should have been
66 "backspace".
67
68 10. Remove two redundant lines of code that can never be obeyed (their function
69 was moved elsewhere).
70
71 11. The program that makes PCRE's Unicode character property table had a bug
72 which caused it to generate incorrect table entries for sequences of
73 characters that have the same character type, but are in different scripts.
74 It amalgamated them into a single range, with the script of the first of
75 them. In other words, some characters were in the wrong script. There were
76 thirteen such cases, affecting characters in the following ranges:
77
78 U+002b0 - U+002c1
79 U+0060c - U+0060d
80 U+0061e - U+00612
81 U+0064b - U+0065e
82 U+0074d - U+0076d
83 U+01800 - U+01805
84 U+01d00 - U+01d77
85 U+01d9b - U+01dbf
86 U+0200b - U+0200f
87 U+030fc - U+030fe
88 U+03260 - U+0327f
89 U+0fb46 - U+0fbb1
90 U+10450 - U+1049d
91
92 12. The -o option (show only the matching part of a line) for pcregrep was not
93 compatible with GNU grep in that, if there was more than one match in a
94 line, it showed only the first of them. It now behaves in the same way as
95 GNU grep.
96
97 13. If the -o and -v options were combined for pcregrep, it printed a blank
98 line for every non-matching line. GNU grep prints nothing, and pcregrep now
99 does the same. The return code can be used to tell if there were any
100 non-matching lines.
101
102 14. Added --file-offsets and --line-offsets to pcregrep.
103
104 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
105 infinitely looping recursion. The bug was that positive lookaheads were not
106 being skipped when checking for a possible empty match (negative lookaheads
107 and both kinds of lookbehind were skipped).
108
109 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
110 inclusion of <windows.h> to before rather than after the definition of
111 INVALID_FILE_ATTRIBUTES (patch from David Byron).
112
113 17. Specifying a possessive quantifier with a specific limit for a Unicode
114 character property caused pcre_compile() to compile bad code, which led at
115 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
116 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
117 caused the error; without that there was no problem.
118
119 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
120
121 19. Added --enable-pcretest-libreadline.
122
123 20. In pcrecpp.cc, the variable 'count' was incremented twice in
124 RE::GlobalReplace(). As a result, the number of replacements returned was
125 double what it should be. I removed one of the increments, but Craig sent a
126 later patch that removed the other one (the right fix) and added unit tests
127 that check the return values (which was not done before).
128
129 21. Several CMake things:
130
131 (1) Arranged that, when cmake is used on Unix, the libraries end up with
132 the names libpcre and libpcreposix, not just pcre and pcreposix.
133
134 (2) The above change means that pcretest and pcregrep are now correctly
135 linked with the newly-built libraries, not previously installed ones.
136
137 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
138
139 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
140 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
141 UTF-8 newline character). The key issue is that the pattern starts .*;
142 this means that the match must be either at the beginning, or after a
143 newline. The bug was in the code for advancing after a failed match and
144 checking that the new position followed a newline. It was not taking
145 account of UTF-8 characters correctly.
146
147 23. PCRE was behaving differently from Perl in the way it recognized POSIX
148 character classes. PCRE was not treating the sequence [:...:] as a
149 character class unless the ... were all letters. Perl, however, seems to
150 allow any characters between [: and :], though of course it rejects as
151 unknown any "names" that contain non-letters, because all the known class
152 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
153 for example, whereas PCRE did not - it did not recognize a POSIX character
154 class. This seemed a bit dangerous, so the code has been changed to be
155 closer to Perl. The behaviour is not identical to Perl, because PCRE will
156 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
157 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
158 Perl does, and where it didn't before.
159
160 24. Rewrite so as to remove the single use of %n from pcregrep because in some
161 Windows environments %n is disabled by default.
162
163
164 Version 7.4 21-Sep-07
165 ---------------------
166
167 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
168 means that a class such as [\s] counted as "explicit reference to CR or
169 LF". That isn't really right - the whole point of the change was to try to
170 help when there was an actual mention of one of the two characters. So now
171 the change happens only if \r or \n (or a literal CR or LF) character is
172 encountered.
173
174 2. The 32-bit options word was also used for 6 internal flags, but the numbers
175 of both had grown to the point where there were only 3 bits left.
176 Fortunately, there was spare space in the data structure, and so I have
177 moved the internal flags into a new 16-bit field to free up more option
178 bits.
179
180 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
181 but did not set the internal JCHANGED flag - either of these is enough to
182 control the way the "get" function works - but the PCRE_INFO_JCHANGED
183 facility is supposed to tell if (?J) was ever used, so now (?J) at the
184 start sets both bits.
185
186 4. Added options (at build time, compile time, exec time) to change \R from
187 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
188
189 5. doc/pcresyntax.html was missing from the distribution.
190
191 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
192 compatibility, even though it is no longer used.
193
194 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
195 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
196 windows.h file is present (where different names are used). [This was
197 reversed later after testing - see 16 below.]
198
199 8. Changed all #include <config.h> to #include "config.h". There were also
200 some further <pcre.h> cases that I changed to "pcre.h".
201
202 9. When pcregrep was used with the --colour option, it missed the line ending
203 sequence off the lines that it output.
204
205 10. It was pointed out to me that arrays of string pointers cause lots of
206 relocations when a shared library is dynamically loaded. A technique of
207 using a single long string with a table of offsets can drastically reduce
208 these. I have refactored PCRE in four places to do this. The result is
209 dramatic:
210
211 Originally: 290
212 After changing UCP table: 187
213 After changing error message table: 43
214 After changing table of "verbs" 36
215 After changing table of Posix names 22
216
217 Thanks to the folks working on Gregex for glib for this insight.
218
219 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
220 unicode-properties was also set.
221
222 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
223
224 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
225 checked only for CRLF.
226
227 14. Added casts to pcretest.c to avoid compiler warnings.
228
229 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
230
231 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
232 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
233 entirely. This removes changes made in 7 above.
234
235 17. The CMake files have been updated, and there is now more information about
236 building with CMake in the NON-UNIX-USE document.
237
238
239 Version 7.3 28-Aug-07
240 ---------------------
241
242 1. In the rejigging of the build system that eventually resulted in 7.1, the
243 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
244 brackets there is not right, since it causes compilers to look for an
245 installed pcre.h, not the version that is in the source that is being
246 compiled (which of course may be different). I have changed it back to:
247
248 #include "pcre.h"
249
250 I have a vague recollection that the change was concerned with compiling in
251 different directories, but in the new build system, that is taken care of
252 by the VPATH setting the Makefile.
253
254 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
255 when the subject happened to end in the byte 0x85 (e.g. if the last
256 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
257 characters but of course it shouldn't be taken as a newline when it is part
258 of another character. The bug was that, for an unlimited repeat of . in
259 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
260 characters when looking for a newline.
261
262 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
263
264 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
265 in debug output.
266
267 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
268 long printing in the pcrecpp unittest when running under MinGW.
269
270 6. ESC_K was left out of the EBCDIC table.
271
272 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
273 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
274 limit also applies to "virtual nesting" when a pattern is recursive, and in
275 this case 1000 isn't so big. I have been able to remove this limit at the
276 expense of backing off one optimization in certain circumstances. Normally,
277 when pcre_exec() would call its internal match() function recursively and
278 immediately return the result unconditionally, it uses a "tail recursion"
279 feature to save stack. However, when a subpattern that can match an empty
280 string has an unlimited repetition quantifier, it no longer makes this
281 optimization. That gives it a stack frame in which to save the data for
282 checking that an empty string has been matched. Previously this was taken
283 from the 1000-entry workspace that had been reserved. So now there is no
284 explicit limit, but more stack is used.
285
286 8. Applied Daniel's patches to solve problems with the import/export magic
287 syntax that is required for Windows, and which was going wrong for the
288 pcreposix and pcrecpp parts of the library. These were overlooked when this
289 problem was solved for the main library.
290
291 9. There were some crude static tests to avoid integer overflow when computing
292 the size of patterns that contain repeated groups with explicit upper
293 limits. As the maximum quantifier is 65535, the maximum group length was
294 set at 30,000 so that the product of these two numbers did not overflow a
295 32-bit integer. However, it turns out that people want to use groups that
296 are longer than 30,000 bytes (though not repeat them that many times).
297 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
298 made it possible to implement the integer overflow checks in a much more
299 dynamic way, which I have now done. The artificial limitation on group
300 length has been removed - we now have only the limit on the total length of
301 the compiled pattern, which depends on the LINK_SIZE setting.
302
303 10. Fixed a bug in the documentation for get/copy named substring when
304 duplicate names are permitted. If none of the named substrings are set, the
305 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
306 empty string.
307
308 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
309 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
310 because the ] is interpreted as the first data character and the
311 terminating ] is not found. PCRE has been made compatible with Perl in this
312 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
313 cause memory overwriting.
314
315 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
316 string has been matched (to stop an infinite loop). It was not recognizing
317 a conditional subpattern that could match an empty string if that
318 subpattern was within another subpattern. For example, it looped when
319 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
320 condition was not nested. This bug has been fixed.
321
322 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
323 past the start of the subject in the presence of bytes with the top bit
324 set, for example "\x8aBCD".
325
326 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
327 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
328
329 14. Optimized (?!) to (*FAIL).
330
331 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
332 This restricts code points to be within the range 0 to 0x10FFFF, excluding
333 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
334 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
335 does: it's just the validity check that is more restrictive.
336
337 16. Inserted checks for integer overflows during escape sequence (backslash)
338 processing, and also fixed erroneous offset values for syntax errors during
339 backslash processing.
340
341 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
342 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
343
344 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
345 caused an overrun.
346
347 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
348 something other than just ASCII characters) inside a group that had an
349 unlimited repeat caused a loop at compile time (while checking to see
350 whether the group could match an empty string).
351
352 20. Debugging a pattern containing \p or \P could cause a crash. For example,
353 [\P{Any}] did so. (Error in the code for printing property names.)
354
355 21. An orphan \E inside a character class could cause a crash.
356
357 22. A repeated capturing bracket such as (A)? could cause a wild memory
358 reference during compilation.
359
360 23. There are several functions in pcre_compile() that scan along a compiled
361 expression for various reasons (e.g. to see if it's fixed length for look
362 behind). There were bugs in these functions when a repeated \p or \P was
363 present in the pattern. These operators have additional parameters compared
364 with \d, etc, and these were not being taken into account when moving along
365 the compiled data. Specifically:
366
367 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
368 length.
369
370 (b) An item such as \pL+ within a repeated group could cause crashes or
371 loops.
372
373 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
374 "reference to non-existent subpattern" error.
375
376 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
377
378 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
379 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
380
381 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
382
383 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
384 character were causing crashes (broken optimization).
385
386 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
387 \p or \P) caused a compile-time loop.
388
389 28. More problems have arisen in unanchored patterns when CRLF is a valid line
390 break. For example, the unstudied pattern [\r\n]A does not match the string
391 "\r\nA" because change 7.0/46 below moves the current point on by two
392 characters after failing to match at the start. However, the pattern \nA
393 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
394 the same is true. There doesn't seem any very clean way out of this, but
395 what I have chosen to do makes the common cases work: PCRE now takes note
396 of whether there can be an explicit match for \r or \n anywhere in the
397 pattern, and if so, 7.0/46 no longer applies. As part of this change,
398 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
399 pattern has explicit CR or LF references.
400
401 29. Added (*CR) etc for changing newline setting at start of pattern.
402
403
404 Version 7.2 19-Jun-07
405 ---------------------
406
407 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
408 which is apparently normally available under Windows.
409
410 2. Re-jig the pcregrep tests with different newline settings in an attempt
411 to make them independent of the local environment's newline setting.
412
413 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
414
415 4. Some of the "internals" tests were previously cut out when the link size
416 was not 2, because the output contained actual offsets. The recent new
417 "Z" feature of pcretest means that these can be cut out, making the tests
418 usable with all link sizes.
419
420 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
421 stack recursion. This gives a massive performance boost under BSD, but just
422 a small improvement under Linux. However, it saves one field in the frame
423 in all cases.
424
425 6. Added more features from the forthcoming Perl 5.10:
426
427 (a) (?-n) (where n is a string of digits) is a relative subroutine or
428 recursion call. It refers to the nth most recently opened parentheses.
429
430 (b) (?+n) is also a relative subroutine call; it refers to the nth next
431 to be opened parentheses.
432
433 (c) Conditions that refer to capturing parentheses can be specified
434 relatively, for example, (?(-2)... or (?(+3)...
435
436 (d) \K resets the start of the current match so that everything before
437 is not part of it.
438
439 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
440
441 (f) \g{name} is another synonym - part of Perl 5.10's unification of
442 reference syntax.
443
444 (g) (?| introduces a group in which the numbering of parentheses in each
445 alternative starts with the same number.
446
447 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
448
449 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
450 PCRE_INFO_JCHANGED.
451
452 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
453 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
454 for detecting groups that can match an empty string.
455
456 9. A pattern with a very large number of alternatives (more than several
457 hundred) was running out of internal workspace during the pre-compile
458 phase, where pcre_compile() figures out how much memory will be needed. A
459 bit of new cunning has reduced the workspace needed for groups with
460 alternatives. The 1000-alternative test pattern now uses 12 bytes of
461 workspace instead of running out of the 4096 that are available.
462
463 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
464
465 11. Applied patch from Google to remove an optimization that didn't quite work.
466 The report of the bug said:
467
468 pcrecpp::RE("a*").FullMatch("aaa") matches, while
469 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
470 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
471
472 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
473 it matched the wrong number of bytes.
474
475
476 Version 7.1 24-Apr-07
477 ---------------------
478
479 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
480 that is more "standard", making use of automake and other Autotools. There
481 is some re-arrangement of the files and adjustment of comments consequent
482 on this.
483
484 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
485 for recursive directory scanning broke on some systems because the files
486 are not scanned in any specific order and on different systems the order
487 was different. A call to "sort" has been inserted into RunGrepTest for the
488 approprate test as a short-term fix. In the longer term there may be an
489 alternative.
490
491 3. I had an email from Eric Raymond about problems translating some of PCRE's
492 man pages to HTML (despite the fact that I distribute HTML pages, some
493 people do their own conversions for various reasons). The problems
494 concerned the use of low-level troff macros .br and .in. I have therefore
495 removed all such uses from the man pages (some were redundant, some could
496 be replaced by .nf/.fi pairs). The 132html script that I use to generate
497 HTML has been updated to handle .nf/.fi and to complain if it encounters
498 .br or .in.
499
500 4. Updated comments in configure.ac that get placed in config.h.in and also
501 arranged for config.h to be included in the distribution, with the name
502 config.h.generic, for the benefit of those who have to compile without
503 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
504
505 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
506 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
507 makevp.bat for the current PCRE, using the additional files
508 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
509
510 6. A Windows user reported a minor discrepancy with test 2, which turned out
511 to be caused by a trailing space on an input line that had got lost in his
512 copy. The trailing space was an accident, so I've just removed it.
513
514 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
515 that is needed.
516
517 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
518 as "const" (a) because they are and (b) because it helps the PHP
519 maintainers who have recently made a script to detect big data structures
520 in the php code that should be moved to the .rodata section. I remembered
521 to update Builducptable as well, so it won't revert if ucptable.h is ever
522 re-created.
523
524 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
525 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
526 order to be able to cut out the UTF-8 tables in the latter when UTF-8
527 support is not required. This saves 1.5-2K of code, which is important in
528 some applications.
529
530 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
531 so as not to refer to the tables, even though these functions will never be
532 called when UTF-8 support is disabled. Otherwise there are problems with a
533 shared library.
534
535 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
536
537 (a) It was defining its arguments as char * instead of void *.
538
539 (b) It was assuming that all moves were upwards in memory; this was true
540 a long time ago when I wrote it, but is no longer the case.
541
542 The emulated memove() is provided for those environments that have neither
543 memmove() nor bcopy(). I didn't think anyone used it these days, but that
544 is clearly not the case, as these two bugs were recently reported.
545
546 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
547 and Detrail to create the HTML documentation, the .txt form of the man
548 pages, and it removes trailing spaces from listed files. It also creates
549 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
550 case, it wraps all the #defines with #ifndefs. This script should be run
551 before "make dist".
552
553 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
554 with Unicode property support.
555
556 (a) For a maximizing quantifier, if the two different cases of the
557 character were of different lengths in their UTF-8 codings (there are
558 some cases like this - I found 11), and the matching function had to
559 back up over a mixture of the two cases, it incorrectly assumed they
560 were both the same length.
561
562 (b) When PCRE was configured to use the heap rather than the stack for
563 recursion during matching, it was not correctly preserving the data for
564 the other case of a UTF-8 character when checking ahead for a match
565 while processing a minimizing repeat. If the check also involved
566 matching a wide character, but failed, corruption could cause an
567 erroneous result when trying to check for a repeat of the original
568 character.
569
570 13. Some tidying changes to the testing mechanism:
571
572 (a) The RunTest script now detects the internal link size and whether there
573 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
574 values substituted by "configure". (The RunGrepTest script already did
575 this for UTF-8.) The configure.ac script no longer substitutes the
576 relevant variables.
577
578 (b) The debugging options /B and /D in pcretest show the compiled bytecode
579 with length and offset values. This means that the output is different
580 for different internal link sizes. Test 2 is skipped for link sizes
581 other than 2 because of this, bypassing the problem. Unfortunately,
582 there was also a test in test 3 (the locale tests) that used /B and
583 failed for link sizes other than 2. Rather than cut the whole test out,
584 I have added a new /Z option to pcretest that replaces the length and
585 offset values with spaces. This is now used to make test 3 independent
586 of link size. (Test 2 will be tidied up later.)
587
588 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
589 segmentation fault instead of returning the appropriate error message.
590
591 15. In multiline mode when the newline sequence was set to "any", the pattern
592 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
593 This doesn't seem right; it now treats the CRLF combination as the line
594 ending, and so does not match in that case. It's only a pattern such as ^$
595 that would hit this one: something like ^ABC$ would have failed after \r
596 and then tried again after \r\n.
597
598 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
599 in an attempt to make files that differ only in their line terminators
600 compare equal. This works on Linux.
601
602 17. Under certain error circumstances pcregrep might try to free random memory
603 as it exited. This is now fixed, thanks to valgrind.
604
605 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
606 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
607 was because its rules for how to advance for /g after matching an empty
608 string at the end of a line did not allow for this case. They now check for
609 it specially.
610
611 20. pcretest is supposed to handle patterns and data of any length, by
612 extending its buffers when necessary. It was getting this wrong when the
613 buffer for a data line had to be extended.
614
615 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
616 CRLF as a newline sequence.
617
618 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
619 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
620 I have nevertheless tidied it up.
621
622 23. Added some casts to kill warnings from HP-UX ia64 compiler.
623
624 24. Added a man page for pcre-config.
625
626
627 Version 7.0 19-Dec-06
628 ---------------------
629
630 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
631 moving to gcc 4.1.1.
632
633 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
634 sys/time.h, which is documented as needed for this function. It doesn't
635 seem to matter on Linux, but it showed up on some releases of OS X.
636
637 3. It seems that there are systems where bytes whose values are greater than
638 127 match isprint() in the "C" locale. The "C" locale should be the
639 default when a C program starts up. In most systems, only ASCII printing
640 characters match isprint(). This difference caused the output from pcretest
641 to vary, making some of the tests fail. I have changed pcretest so that:
642
643 (a) When it is outputting text in the compiled version of a pattern, bytes
644 other than 32-126 are always shown as hex escapes.
645
646 (b) When it is outputting text that is a matched part of a subject string,
647 it does the same, unless a different locale has been set for the match
648 (using the /L modifier). In this case, it uses isprint() to decide.
649
650 4. Fixed a major bug that caused incorrect computation of the amount of memory
651 required for a compiled pattern when options that changed within the
652 pattern affected the logic of the preliminary scan that determines the
653 length. The relevant options are -x, and -i in UTF-8 mode. The result was
654 that the computed length was too small. The symptoms of this bug were
655 either the PCRE error "internal error: code overflow" from pcre_compile(),
656 or a glibc crash with a message such as "pcretest: free(): invalid next
657 size (fast)". Examples of patterns that provoked this bug (shown in
658 pcretest format) are:
659
660 /(?-x: )/x
661 /(?x)(?-x: \s*#\s*)/
662 /((?i)[\x{c0}])/8
663 /(?i:[\x{c0}])/8
664
665 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
666 is now done differently.
667
668 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
669 wrapper classes; (b) implement a new function in the C++ scanner that is
670 more efficient than the old way of doing things because it avoids levels of
671 recursion in the regex matching; (c) add a paragraph to the documentation
672 for the FullMatch() function.
673
674 6. The escape sequence \n was being treated as whatever was defined as
675 "newline". Not only was this contrary to the documentation, which states
676 that \n is character 10 (hex 0A), but it also went horribly wrong when
677 "newline" was defined as CRLF. This has been fixed.
678
679 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
680 was being set to -1 for the "end of line" case (supposedly a value that no
681 character can have). Though this value is never used (the check for end of
682 line is "zero bytes in current character"), it caused compiler complaints.
683 I've changed it to 0xffffffff.
684
685 8. In pcre_version.c, the version string was being built by a sequence of
686 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
687 string (as it is for production releases) called a macro with an empty
688 argument. The C standard says the result of this is undefined. The gcc
689 compiler treats it as an empty string (which was what was wanted) but it is
690 reported that Visual C gives an error. The source has been hacked around to
691 avoid this problem.
692
693 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
694 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
695 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
696 of them did).
697
698 10. Originally, pcretest opened its input and output without "b"; then I was
699 told that "b" was needed in some environments, so it was added for release
700 5.0 to both the input and output. (It makes no difference on Unix-like
701 systems.) Later I was told that it is wrong for the input on Windows. I've
702 now abstracted the modes into two macros, to make it easier to fiddle with
703 them, and removed "b" from the input mode under Windows.
704
705 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
706
707 12. Added -help and --help to pcretest as an official way of being reminded
708 of the options.
709
710 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
711 and pcrecpp.cc because they annoy compilers at high warning levels.
712
713 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
714
715 15. Fixed an occurrence of == in configure.ac that should have been = (shell
716 scripts are not C programs :-) and which was not noticed because it works
717 on Linux.
718
719 16. pcretest is supposed to handle any length of pattern and data line (as one
720 line or as a continued sequence of lines) by extending its input buffer if
721 necessary. This feature was broken for very long pattern lines, leading to
722 a string of junk being passed to pcre_compile() if the pattern was longer
723 than about 50K.
724
725 17. I have done a major re-factoring of the way pcre_compile() computes the
726 amount of memory needed for a compiled pattern. Previously, there was code
727 that made a preliminary scan of the pattern in order to do this. That was
728 OK when PCRE was new, but as the facilities have expanded, it has become
729 harder and harder to keep it in step with the real compile phase, and there
730 have been a number of bugs (see for example, 4 above). I have now found a
731 cunning way of running the real compile function in a "fake" mode that
732 enables it to compute how much memory it would need, while actually only
733 ever using a few hundred bytes of working memory and without too many
734 tests of the mode. This should make future maintenance and development
735 easier. A side effect of this work is that the limit of 200 on the nesting
736 depth of parentheses has been removed (though this was never a serious
737 limitation, I suspect). However, there is a downside: pcre_compile() now
738 runs more slowly than before (30% or more, depending on the pattern). I
739 hope this isn't a big issue. There is no effect on runtime performance.
740
741 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
742 newline (only possible for the last line of a file) and it was a
743 pattern that set a locale (followed by /Lsomething), pcretest crashed.
744
745 19. Added additional timing features to pcretest. (1) The -tm option now times
746 matching only, not compiling. (2) Both -t and -tm can be followed, as a
747 separate command line item, by a number that specifies the number of
748 repeats to use when timing. The default is 50000; this gives better
749 precision, but takes uncomfortably long for very large patterns.
750
751 20. Extended pcre_study() to be more clever in cases where a branch of a
752 subpattern has no definite first character. For example, (a*|b*)[cd] would
753 previously give no result from pcre_study(). Now it recognizes that the
754 first character must be a, b, c, or d.
755
756 21. There was an incorrect error "recursive call could loop indefinitely" if
757 a subpattern (or the entire pattern) that was being tested for matching an
758 empty string contained only one non-empty item after a nested subpattern.
759 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
760 incorrectly, because the \d was being skipped in the check.
761
762 22. The pcretest program now has a new pattern option /B and a command line
763 option -b, which is equivalent to adding /B to every pattern. This causes
764 it to show the compiled bytecode, without the additional information that
765 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
766 is the same as /B/I).
767
768 23. A new optimization is now able automatically to treat some sequences such
769 as a*b as a*+b. More specifically, if something simple (such as a character
770 or a simple class like \d) has an unlimited quantifier, and is followed by
771 something that cannot possibly match the quantified thing, the quantifier
772 is automatically "possessified".
773
774 24. A recursive reference to a subpattern whose number was greater than 39
775 went wrong under certain circumstances in UTF-8 mode. This bug could also
776 have affected the operation of pcre_study().
777
778 25. Realized that a little bit of performance could be had by replacing
779 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
780
781 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
782
783 27. Possessive quantifiers such as a++ were previously implemented by turning
784 them into atomic groups such as ($>a+). Now they have their own opcodes,
785 which improves performance. This includes the automatically created ones
786 from 23 above.
787
788 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
789 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
790 the first matched character to be a colon. This applied both to named and
791 numbered groups.
792
793 29. The ucpinternal.h header file was missing its idempotency #ifdef.
794
795 30. I was sent a "project" file called libpcre.a.dev which I understand makes
796 building PCRE on Windows easier, so I have included it in the distribution.
797
798 31. There is now a check in pcretest against a ridiculously large number being
799 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
800 loop, the loop is abandoned.
801
802 32. Forward references to subpatterns in conditions such as (?(2)...) where
803 subpattern 2 is defined later cause pcre_compile() to search forwards in
804 the pattern for the relevant set of parentheses. This search went wrong
805 when there were unescaped parentheses in a character class, parentheses
806 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
807
808 33. "Subroutine" calls and backreferences were previously restricted to
809 referencing subpatterns earlier in the regex. This restriction has now
810 been removed.
811
812 34. Added a number of extra features that are going to be in Perl 5.10. On the
813 whole, these are just syntactic alternatives for features that PCRE had
814 previously implemented using the Python syntax or my own invention. The
815 other formats are all retained for compatibility.
816
817 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
818 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
819 also .NET compatible.
820
821 (b) A recursion or subroutine call to a named group can now be defined as
822 (?&name) as well as (?P>name).
823
824 (c) A backreference to a named group can now be defined as \k<name> or
825 \k'name' as well as (?P=name). The new forms, as well as being in Perl
826 5.10, are also .NET compatible.
827
828 (d) A conditional reference to a named group can now use the syntax
829 (?(<name>) or (?('name') as well as (?(name).
830
831 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
832 groups (named and numbered) that are never evaluated inline, but can be
833 called as "subroutines" from elsewhere. In effect, the DEFINE condition
834 is always false. There may be only one alternative in such a group.
835
836 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
837 as the simple (?(R). The condition is true only if the most recent
838 recursion is that of the given number or name. It does not search out
839 through the entire recursion stack.
840
841 (g) The escape \gN or \g{N} has been added, where N is a positive or
842 negative number, specifying an absolute or relative reference.
843
844 35. Tidied to get rid of some further signed/unsigned compiler warnings and
845 some "unreachable code" warnings.
846
847 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
848 things, this adds five new scripts.
849
850 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
851 There were also incompatibilities regarding the handling of \Q..\E inside
852 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
853 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
854
855 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
856 matches an empty string, and forcibly breaks the loop. There were bugs in
857 this code in non-simple cases. For a pattern such as ^(a()*)* matched
858 against aaaa the result was just "a" rather than "aaaa", for example. Two
859 separate and independent bugs (that affected different cases) have been
860 fixed.
861
862 39. Refactored the code to abolish the use of different opcodes for small
863 capturing bracket numbers. This is a tidy that I avoided doing when I
864 removed the limit on the number of capturing brackets for 3.5 back in 2001.
865 The new approach is not only tidier, it makes it possible to reduce the
866 memory needed to fix the previous bug (38).
867
868 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
869 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
870 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
871 mode.
872
873 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
874 report.
875
876 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
877 copy construction and assignment in the C++ wrapper.
878
879 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
880 couple of bugs that could have given wrong results in the "--newline=crlf"
881 case.
882
883 44. Added a number of casts and did some reorganization of signed/unsigned int
884 variables following suggestions from Dair Grant. Also renamed the variable
885 "this" as "item" because it is a C++ keyword.
886
887 45. Arranged for dftables to add
888
889 #include "pcre_internal.h"
890
891 to pcre_chartables.c because without it, gcc 4.x may remove the array
892 definition from the final binary if PCRE is built into a static library and
893 dead code stripping is activated.
894
895 46. For an unanchored pattern, if a match attempt fails at the start of a
896 newline sequence, and the newline setting is CRLF or ANY, and the next two
897 characters are CRLF, advance by two characters instead of one.
898
899
900 Version 6.7 04-Jul-06
901 ---------------------
902
903 1. In order to handle tests when input lines are enormously long, pcretest has
904 been re-factored so that it automatically extends its buffers when
905 necessary. The code is crude, but this _is_ just a test program. The
906 default size has been increased from 32K to 50K.
907
908 2. The code in pcre_study() was using the value of the re argument before
909 testing it for NULL. (Of course, in any sensible call of the function, it
910 won't be NULL.)
911
912 3. The memmove() emulation function in pcre_internal.h, which is used on
913 systems that lack both memmove() and bcopy() - that is, hardly ever -
914 was missing a "static" storage class specifier.
915
916 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
917 containing an extended class (one that cannot be represented by a bitmap
918 because it contains high-valued characters or Unicode property items, e.g.
919 [\pZ]). Almost always one would set UTF-8 mode when processing such a
920 pattern, but PCRE should not loop if you do not (it no longer does).
921 [Detail: two cases were found: (a) a repeated subpattern containing an
922 extended class; (b) a recursive reference to a subpattern that followed a
923 previous extended class. It wasn't skipping over the extended class
924 correctly when UTF-8 mode was not set.]
925
926 5. A negated single-character class was not being recognized as fixed-length
927 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
928 compile error "lookbehind assertion is not fixed length".
929
930 6. The RunPerlTest auxiliary script was showing an unexpected difference
931 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
932 write a Perl script that can interpret lines of an input file either as
933 byte characters or as UTF-8, which is what "perltest" was being required to
934 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
935 can't do is switch easily at run time between having the "use utf8;" pragma
936 or not. In the end, I fudged it by using the RunPerlTest script to insert
937 "use utf8;" explicitly for the UTF-8 tests.
938
939 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
940 the end of the subject string, contrary to the documentation and to what
941 Perl does. This was true of both matching functions. Now it matches only at
942 the start of the subject and immediately after *internal* newlines.
943
944 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
945 a pointer to an int instead of a pointer to an unsigned long int. This
946 caused problems on 64-bit systems.
947
948 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
949 instance of the 'standard' template library not being so standard".
950
951 10. There was no check on the number of named subpatterns nor the maximum
952 length of a subpattern name. The product of these values is used to compute
953 the size of the memory block for a compiled pattern. By supplying a very
954 long subpattern name and a large number of named subpatterns, the size
955 computation could be caused to overflow. This is now prevented by limiting
956 the length of names to 32 characters, and the number of named subpatterns
957 to 10,000.
958
959 11. Subpatterns that are repeated with specific counts have to be replicated in
960 the compiled pattern. The size of memory for this was computed from the
961 length of the subpattern and the repeat count. The latter is limited to
962 65535, but there was no limit on the former, meaning that integer overflow
963 could in principle occur. The compiled length of a repeated subpattern is
964 now limited to 30,000 bytes in order to prevent this.
965
966 12. Added the optional facility to have named substrings with the same name.
967
968 13. Added the ability to use a named substring as a condition, using the
969 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
970 are numbers (not recommended). Forward references are permitted.
971
972 14. Added forward references in named backreferences (if you see what I mean).
973
974 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
975 pattern could run off the end of the subject. For example, the pattern
976 "(?s)(.{1,5})"8 did this with the subject "ab".
977
978 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
979 PCRE_CASELESS was set when matching characters that were quantified with ?
980 or *.
981
982 17. A character class other than a single negated character that had a minimum
983 but no maximum quantifier - for example [ab]{6,} - was not handled
984 correctly by pce_dfa_exec(). It would match only one character.
985
986 18. A valid (though odd) pattern that looked like a POSIX character
987 class but used an invalid character after [ (for example [[,abc,]]) caused
988 pcre_compile() to give the error "Failed: internal error: code overflow" or
989 in some cases to crash with a glibc free() error. This could even happen if
990 the pattern terminated after [[ but there just happened to be a sequence of
991 letters, a binary zero, and a closing ] in the memory that followed.
992
993 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
994 over the years. Originally (before any Unicode support), just the bottom 8
995 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
996 output from "man perlunicode" includes this:
997
998 The regular expression compiler produces polymorphic opcodes. That
999 is, the pattern adapts to the data and automatically switches to
1000 the Unicode character scheme when presented with Unicode data--or
1001 instead uses a traditional byte scheme when presented with byte
1002 data.
1003
1004 Sadly, a wide octal escape does not cause a switch, and in a string with
1005 no other multibyte characters, these octal escapes are treated as before.
1006 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1007 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1008 Unicode string.
1009
1010 I have not perpetrated such confusion in PCRE. Up till now, it took just
1011 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1012 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1013 translate to the appropriate multibyte character.
1014
1015 29. Applied some refactoring to reduce the number of warnings from Microsoft
1016 and Borland compilers. This has included removing the fudge introduced
1017 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1018 a warning about an unused variable.
1019
1020 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1021 characters since release 4.0, because Perl (from release 5.004) does not.
1022 [Or at least, is documented not to: some releases seem to be in conflict
1023 with the documentation.] However, when a pattern was studied with
1024 pcre_study() and all its branches started with \s, PCRE still included VT
1025 as a possible starting character. Of course, this did no harm; it just
1026 caused an unnecessary match attempt.
1027
1028 22. Removed a now-redundant internal flag bit that recorded the fact that case
1029 dependency changed within the pattern. This was once needed for "required
1030 byte" processing, but is no longer used. This recovers a now-scarce options
1031 bit. Also moved the least significant internal flag bit to the most-
1032 significant bit of the word, which was not previously used (hangover from
1033 the days when it was an int rather than a uint) to free up another bit for
1034 the future.
1035
1036 23. Added support for CRLF line endings as well as CR and LF. As well as the
1037 default being selectable at build time, it can now be changed at runtime
1038 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1039 specify that it is scanning data with non-default line endings.
1040
1041 24. Changed the definition of CXXLINK to make it agree with the definition of
1042 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1043
1044 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1045 recursions. This makes a big different to stack usage for some patterns.
1046
1047 26. If a subpattern containing a named recursion or subroutine reference such
1048 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1049 the space required for the compiled pattern went wrong and gave too small a
1050 value. Depending on the environment, this could lead to "Failed: internal
1051 error: code overflow at offset 49" or "glibc detected double free or
1052 corruption" errors.
1053
1054 27. Applied patches from Google (a) to support the new newline modes and (b) to
1055 advance over multibyte UTF-8 characters in GlobalReplace.
1056
1057 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1058 difference for some implementation of PCRE in some Windows version.
1059
1060 29. Added some extra testing facilities to pcretest:
1061
1062 \q<number> in a data line sets the "match limit" value
1063 \Q<number> in a data line sets the "match recursion limt" value
1064 -S <number> sets the stack size, where <number> is in megabytes
1065
1066 The -S option isn't available for Windows.
1067
1068
1069 Version 6.6 06-Feb-06
1070 ---------------------
1071
1072 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1073 in pcreposix.h. I have copied the definition from pcre.h.
1074
1075 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1076 because pcre.h is no longer a built file.
1077
1078 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1079 not normally included in the compiled code.
1080
1081
1082 Version 6.5 01-Feb-06
1083 ---------------------
1084
1085 1. When using the partial match feature with pcre_dfa_exec(), it was not
1086 anchoring the second and subsequent partial matches at the new starting
1087 point. This could lead to incorrect results. For example, with the pattern
1088 /1234/, partially matching against "123" and then "a4" gave a match.
1089
1090 2. Changes to pcregrep:
1091
1092 (a) All non-match returns from pcre_exec() were being treated as failures
1093 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1094 error message is output. Some extra information is given for the
1095 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1096 probably the only errors that are likely to be caused by users (by
1097 specifying a regex that has nested indefinite repeats, for instance).
1098 If there are more than 20 of these errors, pcregrep is abandoned.
1099
1100 (b) A binary zero was treated as data while matching, but terminated the
1101 output line if it was written out. This has been fixed: binary zeroes
1102 are now no different to any other data bytes.
1103
1104 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1105 used to set a locale for matching. The --locale=xxxx long option has
1106 been added (no short equivalent) to specify a locale explicitly on the
1107 pcregrep command, overriding the environment variables.
1108
1109 (d) When -B was used with -n, some line numbers in the output were one less
1110 than they should have been.
1111
1112 (e) Added the -o (--only-matching) option.
1113
1114 (f) If -A or -C was used with -c (count only), some lines of context were
1115 accidentally printed for the final match.
1116
1117 (g) Added the -H (--with-filename) option.
1118
1119 (h) The combination of options -rh failed to suppress file names for files
1120 that were found from directory arguments.
1121
1122 (i) Added the -D (--devices) and -d (--directories) options.
1123
1124 (j) Added the -F (--fixed-strings) option.
1125
1126 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1127
1128 (l) Added the --colo(u)r option.
1129
1130 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1131 is not present by default.
1132
1133 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1134 items such as (?R) or (?1), when the recursion could match a number of
1135 alternatives. If it matched one of the alternatives, but subsequently,
1136 outside the recursion, there was a failure, the code tried to back up into
1137 the recursion. However, because of the way PCRE is implemented, this is not
1138 possible, and the result was an incorrect result from the match.
1139
1140 In order to prevent this happening, the specification of recursion has
1141 been changed so that all such subpatterns are automatically treated as
1142 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1143
1144 4. I had overlooked the fact that, in some locales, there are characters for
1145 which isalpha() is true but neither isupper() nor islower() are true. In
1146 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1147 and ordfeminine) are like this. This affected the treatment of \w and \W
1148 when they appeared in character classes, but not when they appeared outside
1149 a character class. The bit map for "word" characters is now created
1150 separately from the results of isalnum() instead of just taking it from the
1151 upper, lower, and digit maps. (Plus the underscore character, of course.)
1152
1153 5. The above bug also affected the handling of POSIX character classes such as
1154 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1155 permanent tables. Instead, the bit maps for such a class were previously
1156 created as the appropriate unions of the upper, lower, and digit bitmaps.
1157 Now they are created by subtraction from the [[:word:]] class, which has
1158 its own bitmap.
1159
1160 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1161 It is created by subtracting the vertical space characters (\x09, \x0a,
1162 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1163 subtraction was done in the overall bitmap for a character class, meaning
1164 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1165 be recognized. This bug has been fixed.
1166
1167 7. Patches from the folks at Google:
1168
1169 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1170 real life, but is still worth protecting against".
1171
1172 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1173 regular expressions".
1174
1175 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1176 have it.
1177
1178 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1179 "configure" and the latter not, in order to fix a problem somebody had
1180 with compiling the Arg class on HP-UX.
1181
1182 (e) Improve the error-handling of the C++ wrapper a little bit.
1183
1184 (f) New tests for checking recursion limiting.
1185
1186 8. The pcre_memmove() function, which is used only if the environment does not
1187 have a standard memmove() function (and is therefore rarely compiled),
1188 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1189 returning a result (though PCRE never actually uses the result).
1190
1191 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1192 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1193 returned instead of calling malloc() with an overflowing number that would
1194 most likely cause subsequent chaos.
1195
1196 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1197
1198 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1199 with this option is matched, the nmatch and pmatch options of regexec() are
1200 ignored.
1201
1202 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1203 provided in case anyone wants to the the POSIX interface with UTF-8
1204 strings.
1205
1206 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1207 C++ linking (needed for some HP-UX environments).
1208
1209 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1210 (unused parameter) and in the pcre_printint() function (omitted "default"
1211 switch label when the default is to do nothing).
1212
1213 15. Added some code to make it possible, when PCRE is compiled as a C++
1214 library, to replace subject pointers for pcre_exec() with a smart pointer
1215 class, thus making it possible to process discontinuous strings.
1216
1217 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1218 much the same function. They were added by different people who were trying
1219 to make PCRE easy to compile on non-Unix systems. It has been suggested
1220 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1221 for compiling on Windows systems. I have therefore replaced it with
1222 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1223 defaults to "extern" for C or "extern C" for C++, which works fine on
1224 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1225 SCOPE with something explicit in config.h. In addition:
1226
1227 (a) pcreposix.h still had just "extern" instead of either of these macros;
1228 I have replaced it with PCRE_DATA_SCOPE.
1229
1230 (b) Functions such as _pcre_xclass(), which are internal to the library,
1231 but external in the C sense, all had PCRE_EXPORT in their definitions.
1232 This is apparently wrong for the Windows case, so I have removed it.
1233 (It makes no difference on Unix-like systems.)
1234
1235 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1236 of recursive calls to match(). This is different to MATCH_LIMIT because
1237 that limits the total number of calls to match(), not all of which increase
1238 the depth of recursion. Limiting the recursion depth limits the amount of
1239 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1240 when PCRE is compiled, and changed at run time. A patch from Google adds
1241 this functionality to the C++ interface.
1242
1243 18. Changes to the handling of Unicode character properties:
1244
1245 (a) Updated the table to Unicode 4.1.0.
1246
1247 (b) Recognize characters that are not in the table as "Cn" (undefined).
1248
1249 (c) I revised the way the table is implemented to a much improved format
1250 which includes recognition of ranges. It now supports the ranges that
1251 are defined in UnicodeData.txt, and it also amalgamates other
1252 characters into ranges. This has reduced the number of entries in the
1253 table from around 16,000 to around 3,000, thus reducing its size
1254 considerably. I realized I did not need to use a tree structure after
1255 all - a binary chop search is just as efficient. Having reduced the
1256 number of entries, I extended their size from 6 bytes to 8 bytes to
1257 allow for more data.
1258
1259 (d) Added support for Unicode script names via properties such as \p{Han}.
1260
1261 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1262 matching that character.
1263
1264 20. When matching a repeated Unicode property with a minimum greater than zero,
1265 (for example \pL{2,}), PCRE could look past the end of the subject if it
1266 reached it while seeking the minimum number of characters. This could
1267 happen only if some of the characters were more than one byte long, because
1268 there is a check for at least the minimum number of bytes.
1269
1270 21. Refactored the implementation of \p and \P so as to be more general, to
1271 allow for more different types of property in future. This has changed the
1272 compiled form incompatibly. Anybody with saved compiled patterns that use
1273 \p or \P will have to recompile them.
1274
1275 22. Added "Any" and "L&" to the supported property types.
1276
1277 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1278 but give a compile time error if the value is greater than 0xff.
1279
1280 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1281 accidentally not being installed or uninstalled.
1282
1283 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1284 made were to insert the current release number. This seemed silly, because
1285 it made things harder for people building PCRE on systems that don't run
1286 "configure". I have turned pcre.h into a distributed file, no longer built
1287 by "configure", with the version identification directly included. There is
1288 no longer a pcre.h.in file.
1289
1290 However, this change necessitated a change to the pcre-config script as
1291 well. It is built from pcre-config.in, and one of the substitutions was the
1292 release number. I have updated configure.ac so that ./configure now finds
1293 the release number by grepping pcre.h.
1294
1295 26. Added the ability to run the tests under valgrind.
1296
1297
1298 Version 6.4 05-Sep-05
1299 ---------------------
1300
1301 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1302 "--" to be printed when multiple files were scanned, even when none of the
1303 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1304 consider it to be a bug, and have restored the previous behaviour.
1305
1306 2. A couple of code tidies to get rid of compiler warnings.
1307
1308 3. The pcretest program used to cheat by referring to symbols in the library
1309 whose names begin with _pcre_. These are internal symbols that are not
1310 really supposed to be visible externally, and in some environments it is
1311 possible to suppress them. The cheating is now confined to including
1312 certain files from the library's source, which is a bit cleaner.
1313
1314 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1315 file's purpose clearer.
1316
1317 5. Reorganized pcre_ucp_findchar().
1318
1319
1320 Version 6.3 15-Aug-05
1321 ---------------------
1322
1323 1. The file libpcre.pc.in did not have general read permission in the tarball.
1324
1325 2. There were some problems when building without C++ support:
1326
1327 (a) If C++ support was not built, "make install" and "make test" still
1328 tried to test it.
1329
1330 (b) There were problems when the value of CXX was explicitly set. Some
1331 changes have been made to try to fix these, and ...
1332
1333 (c) --disable-cpp can now be used to explicitly disable C++ support.
1334
1335 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1336 backslash in a target when C++ was disabled. This confuses some
1337 versions of "make", apparently. Using an intermediate variable solves
1338 this. (Same for CPP_LOBJ.)
1339
1340 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1341 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1342 necessary on certain architectures.
1343
1344 4. Added a setting of -export-symbols-regex to the link command to remove
1345 those symbols that are exported in the C sense, but actually are local
1346 within the library, and not documented. Their names all begin with
1347 "_pcre_". This is not a perfect job, because (a) we have to except some
1348 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1349 available (and never for static libraries). I have made a note to try to
1350 find a way round (a) in the future.
1351
1352
1353 Version 6.2 01-Aug-05
1354 ---------------------
1355
1356 1. There was no test for integer overflow of quantifier values. A construction
1357 such as {1111111111111111} would give undefined results. What is worse, if
1358 a minimum quantifier for a parenthesized subpattern overflowed and became
1359 negative, the calculation of the memory size went wrong. This could have
1360 led to memory overwriting.
1361
1362 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1363
1364 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1365 operating environments where this matters.
1366
1367 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1368 PCRE options from within the C++ wrapper.
1369
1370 5. Named capturing subpatterns were not being correctly counted when a pattern
1371 was compiled. This caused two problems: (a) If there were more than 100
1372 such subpatterns, the calculation of the memory needed for the whole
1373 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1374 back references of the form \12, where the number was greater than 9, were
1375 not recognized as back references, even though there were sufficient
1376 previous subpatterns.
1377
1378 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1379 versions of gcc, e.g. 2.95.4.
1380
1381
1382 Version 6.1 21-Jun-05
1383 ---------------------
1384
1385 1. There was one reference to the variable "posix" in pcretest.c that was not
1386 surrounded by "#if !defined NOPOSIX".
1387
1388 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1389 the cross-check on the old pcre_info() function, for the benefit of the
1390 cut-down version of PCRE that is currently imported into Exim.
1391
1392 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1393 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1394 patterns that start (?i) but I don't think that matters. The use of (?i) is
1395 just an example; this all applies to the other options as well.
1396
1397 4. Since libtool seems to echo the compile commands it is issuing, the output
1398 from "make" can be reduced a bit by putting "@" in front of each libtool
1399 compile command.
1400
1401 5. Patch from the folks at Google for configure.in to be a bit more thorough
1402 in checking for a suitable C++ installation before trying to compile the
1403 C++ stuff. This should fix a reported problem when a compiler was present,
1404 but no suitable headers.
1405
1406 6. The man pages all had just "PCRE" as their title. I have changed them to
1407 be the relevant file name. I have also arranged that these names are
1408 retained in the file doc/pcre.txt, which is a concatenation in text format
1409 of all the man pages except the little individual ones for each function.
1410
1411 7. The NON-UNIX-USE file had not been updated for the different set of source
1412 files that come with release 6. I also added a few comments about the C++
1413 wrapper.
1414
1415
1416 Version 6.0 07-Jun-05
1417 ---------------------
1418
1419 1. Some minor internal re-organization to help with my DFA experiments.
1420
1421 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1422 didn't matter for the library itself when fully configured, but did matter
1423 when compiling without UCP support, or within Exim, where the ucp files are
1424 not imported.
1425
1426 3. Refactoring of the library code to split up the various functions into
1427 different source modules. The addition of the new DFA matching code (see
1428 below) to a single monolithic source would have made it really too
1429 unwieldy, quite apart from causing all the code to be include in a
1430 statically linked application, when only some functions are used. This is
1431 relevant even without the DFA addition now that patterns can be compiled in
1432 one application and matched in another.
1433
1434 The downside of splitting up is that there have to be some external
1435 functions and data tables that are used internally in different modules of
1436 the library but which are not part of the API. These have all had their
1437 names changed to start with "_pcre_" so that they are unlikely to clash
1438 with other external names.
1439
1440 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1441 a different (DFA) algorithm. Although it is slower than the original
1442 function, it does have some advantages for certain types of matching
1443 problem.
1444
1445 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1446 including restarting after a partial match.
1447
1448 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1449 defined when compiling for Windows was sent to me. I have put it into the
1450 code, though I have no means of testing or verifying it.
1451
1452 7. Added the pcre_refcount() auxiliary function.
1453
1454 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1455 match before or at the first newline in the subject string. In pcretest,
1456 the /f option on a pattern can be used to set this.
1457
1458 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1459 would behave wrongly. This has been present in PCRE since release 4.0.
1460
1461 10. A number of changes to the pcregrep command:
1462
1463 (a) Refactored how -x works; insert ^(...)$ instead of setting
1464 PCRE_ANCHORED and checking the length, in preparation for adding
1465 something similar for -w.
1466
1467 (b) Added the -w (match as a word) option.
1468
1469 (c) Refactored the way lines are read and buffered so as to have more
1470 than one at a time available.
1471
1472 (d) Implemented a pcregrep test script.
1473
1474 (e) Added the -M (multiline match) option. This allows patterns to match
1475 over several lines of the subject. The buffering ensures that at least
1476 8K, or the rest of the document (whichever is the shorter) is available
1477 for matching (and similarly the previous 8K for lookbehind assertions).
1478
1479 (f) Changed the --help output so that it now says
1480
1481 -w, --word-regex(p)
1482
1483 instead of two lines, one with "regex" and the other with "regexp"
1484 because that confused at least one person since the short forms are the
1485 same. (This required a bit of code, as the output is generated
1486 automatically from a table. It wasn't just a text change.)
1487
1488 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1489 option but starts with a hyphen. Could be a pattern or a path name
1490 starting with a hyphen, for instance.
1491
1492 (h) "-" can be given as a file name to represent stdin.
1493
1494 (i) When file names are being printed, "(standard input)" is used for
1495 the standard input, for compatibility with GNU grep. Previously
1496 "<stdin>" was used.
1497
1498 (j) The option --label=xxx can be used to supply a name to be used for
1499 stdin when file names are being printed. There is no short form.
1500
1501 (k) Re-factored the options decoding logic because we are going to add
1502 two more options that take data. Such options can now be given in four
1503 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1504
1505 (l) Added the -A, -B, and -C options for requesting that lines of context
1506 around matches be printed.
1507
1508 (m) Added the -L option to print the names of files that do not contain
1509 any matching lines, that is, the complement of -l.
1510
1511 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1512 continue to scan other files.
1513
1514 (o) The -s option was incorrectly implemented. For compatibility with other
1515 greps, it now suppresses the error message for a non-existent or non-
1516 accessible file (but not the return code). There is a new option called
1517 -q that suppresses the output of matching lines, which was what -s was
1518 previously doing.
1519
1520 (p) Added --include and --exclude options to specify files for inclusion
1521 and exclusion when recursing.
1522
1523 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1524 Hopefully, it now does.
1525
1526 12. Missing cast in pcre_study().
1527
1528 13. Added an "uninstall" target to the makefile.
1529
1530 14. Replaced "extern" in the function prototypes in Makefile.in with
1531 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1532 world, but is set differently for Windows.
1533
1534 15. Added a second compiling function called pcre_compile2(). The only
1535 difference is that it has an extra argument, which is a pointer to an
1536 integer error code. When there is a compile-time failure, this is set
1537 non-zero, in addition to the error test pointer being set to point to an
1538 error message. The new argument may be NULL if no error number is required
1539 (but then you may as well call pcre_compile(), which is now just a
1540 wrapper). This facility is provided because some applications need a
1541 numeric error indication, but it has also enabled me to tidy up the way
1542 compile-time errors are handled in the POSIX wrapper.
1543
1544 16. Added VPATH=.libs to the makefile; this should help when building with one
1545 prefix path and installing with another. (Or so I'm told by someone who
1546 knows more about this stuff than I do.)
1547
1548 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1549 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1550 match everything, including newlines. This is not POSIX-compatible, but
1551 somebody wanted the feature. From pcretest it can be activated by using
1552 both the P and the s flags.
1553
1554 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1555
1556 19. libpcre.pc was being incorrectly installed as executable.
1557
1558 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1559 it now also looks for '\r' so that it will work unmodified on Windows.
1560
1561 21. Added Google's contributed C++ wrapper to the distribution.
1562
1563 22. Added some untidy missing memory free() calls in pcretest, to keep
1564 Electric Fence happy when testing.
1565
1566
1567
1568 Version 5.0 13-Sep-04
1569 ---------------------
1570
1571 1. Internal change: literal characters are no longer packed up into items
1572 containing multiple characters in a single byte-string. Each character
1573 is now matched using a separate opcode. However, there may be more than one
1574 byte in the character in UTF-8 mode.
1575
1576 2. The pcre_callout_block structure has two new fields: pattern_position and
1577 next_item_length. These contain the offset in the pattern to the next match
1578 item, and its length, respectively.
1579
1580 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1581 insertion of callouts before each pattern item. Added the /C option to
1582 pcretest to make use of this.
1583
1584 4. On the advice of a Windows user, the lines
1585
1586 #if defined(_WIN32) || defined(WIN32)
1587 _setmode( _fileno( stdout ), 0x8000 );
1588 #endif /* defined(_WIN32) || defined(WIN32) */
1589
1590 have been added to the source of pcretest. This apparently does useful
1591 magic in relation to line terminators.
1592
1593 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1594 for the benefit of those environments where the "b" makes a difference.
1595
1596 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1597 to know about it. I have put a hack into configure.in that adds in code
1598 to set GCC=yes if CC=icc. This seems to end up at a point in the
1599 generated configure script that is early enough to affect the setting of
1600 compiler options, which is what is needed, but I have no means of testing
1601 whether it really works. (The user who reported this had patched the
1602 generated configure script, which of course I cannot do.)
1603
1604 LATER: After change 22 below (new libtool files), the configure script
1605 seems to know about icc (and also ecc). Therefore, I have commented out
1606 this hack in configure.in.
1607
1608 7. Added support for pkg-config (2 patches were sent in).
1609
1610 8. Negated POSIX character classes that used a combination of internal tables
1611 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1612 [[:^ascii]]. Typically, they would match almost any characters. The other
1613 POSIX classes were not broken in this way.
1614
1615 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1616 to find the match, as PCRE was deluded into thinking that the match had to
1617 start at the start point or following a newline. The same bug applied to
1618 patterns with negative forward assertions or any backward assertions
1619 preceding ".*" at the start, unless the pattern required a fixed first
1620 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1621
1622 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1623 starting at the last subject character, bytes beyond the end of the subject
1624 string were read.
1625
1626 11. Renamed the variable "class" as "classbits" to make life easier for C++
1627 users. (Previously there was a macro definition, but it apparently wasn't
1628 enough.)
1629
1630 12. Added the new field "tables" to the extra data so that tables can be passed
1631 in at exec time, or the internal tables can be re-selected. This allows
1632 a compiled regex to be saved and re-used at a later time by a different
1633 program that might have everything at different addresses.
1634
1635 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1636 -R library as well as a -L library.
1637
1638 14. The debugging options of pcretest (-d on the command line or D on a
1639 pattern) showed incorrect output for anything following an extended class
1640 that contained multibyte characters and which was followed by a quantifier.
1641
1642 15. Added optional support for general category Unicode character properties
1643 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1644 support. It adds about 90K to the size of the library. The meanings of the
1645 inbuilt class escapes such as \d and \s have NOT been changed.
1646
1647 16. Updated pcredemo.c to include calls to free() to release the memory for the
1648 compiled pattern.
1649
1650 17. The generated file chartables.c was being created in the source directory
1651 instead of in the building directory. This caused the build to fail if the
1652 source directory was different from the building directory, and was
1653 read-only.
1654
1655 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1656 file. No doubt somebody will tell me if they don't make sense... Also added
1657 Dan Mooney's comments about building on OpenVMS.
1658
1659 19. Added support for partial matching via the PCRE_PARTIAL option for
1660 pcre_exec() and the \P data escape in pcretest.
1661
1662 20. Extended pcretest with 3 new pattern features:
1663
1664 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1665 write the compiled pattern to the file whose name is "rest-of-line".
1666 This is a straight binary dump of the data, with the saved pointer to
1667 the character tables forced to be NULL. The study data, if any, is
1668 written too. After writing, pcretest reads a new pattern.
1669
1670 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1671 compiled pattern from the given file. There must not be any
1672 occurrences of "<" in the file name (pretty unlikely); if there are,
1673 pcretest will instead treat the initial "<" as a pattern delimiter.
1674 After reading in the pattern, pcretest goes on to read data lines as
1675 usual.
1676
1677 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1678 and 16-bit fields in a compiled pattern, to simulate a pattern that
1679 was compiled on a host of opposite endianness.
1680
1681 21. The pcre-exec() function can now cope with patterns that were compiled on
1682 hosts of opposite endianness, with this restriction:
1683
1684 As for any compiled expression that is saved and used later, the tables
1685 pointer field cannot be preserved; the extra_data field in the arguments
1686 to pcre_exec() should be used to pass in a tables address if a value
1687 other than the default internal tables were used at compile time.
1688
1689 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1690 now diagnosed as an error. Previously, most of the time, a negative number
1691 would have been treated as zero, but if in addition "ovector" was passed as
1692 NULL, a crash could occur.
1693
1694 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1695 new versions from the libtool 1.5 distribution (the last one is a copy of
1696 a file called libtool.m4). This seems to have fixed the need to patch
1697 "configure" to support Darwin 1.3 (which I used to do). However, I still
1698 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1699 workstation).
1700
1701 24. Changed the PCRE licence to be the more standard "BSD" licence.
1702
1703
1704 Version 4.5 01-Dec-03
1705 ---------------------
1706
1707 1. There has been some re-arrangement of the code for the match() function so
1708 that it can be compiled in a version that does not call itself recursively.
1709 Instead, it keeps those local variables that need separate instances for
1710 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1711 needs to "recurse". Keeping track of where control must go is done by means
1712 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1713 hide most of the details from the main code, and operates only if
1714 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1715 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1716 operating.
1717
1718 To make it easier for callers to provide specially tailored get/free
1719 functions for this usage, two new functions, pcre_stack_malloc, and
1720 pcre_stack_free, are used. They are always called in strict stacking order,
1721 and the size of block requested is always the same.
1722
1723 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1724 PCRE has been compiled to use the stack or the heap for recursion. The
1725 -C option of pcretest uses this to show which version is compiled.
1726
1727 A new data escape \S, is added to pcretest; it causes the amounts of store
1728 obtained and freed by both kinds of malloc/free at match time to be added
1729 to the output.
1730
1731 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1732 what's available on my current Linux desktop machine.
1733
1734 3. When matching a UTF-8 string, the test for a valid string at the start has
1735 been extended. If start_offset is not zero, PCRE now checks that it points
1736 to a byte that is the start of a UTF-8 character. If not, it returns
1737 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1738 this is necessary because there may be backward assertions in the pattern.
1739 When matching the same subject several times, it may save resources to use
1740 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1741
1742 4. The code for checking the validity of UTF-8 strings has been tightened so
1743 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1744 containing "overlong sequences".
1745
1746 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1747 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1748 should have been "&", but it just so happened that all the cases this let
1749 through by mistake were picked up later in the function.
1750
1751 6. I had used a variable called "isblank" - this is a C99 function, causing
1752 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1753
1754 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1755 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1756 so the version is shown; (c) stop "make test" echoing "./RunTest".
1757
1758 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1759
1760 9. The prototype for memmove() for systems that don't have it was using
1761 size_t, but the inclusion of the header that defines size_t was later. I've
1762 moved the #includes for the C headers earlier to avoid this.
1763
1764 10. Added some adjustments to the code to make it easier to compiler on certain
1765 special systems:
1766
1767 (a) Some "const" qualifiers were missing.
1768 (b) Added the macro EXPORT before all exported functions; by default this
1769 is defined to be empty.
1770 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1771 that it reads its output file name as an argument instead of writing
1772 to the standard output and assuming this can be redirected.
1773
1774 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1775 class containing characters with values greater than 255, PCRE compilation
1776 went into a loop.
1777
1778 12. A recursive reference to a subpattern that was within another subpattern
1779 that had a minimum quantifier of zero caused PCRE to crash. For example,
1780 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1781 recursion. If the recursively-called subpattern itself had a zero repeat,
1782 that was OK.
1783
1784 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1785 buffer into which it was copied (for escape processing) was still set at
1786 1024, so long lines caused crashes.
1787
1788 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1789 "internal error: code overflow...". This applied to any character class
1790 that was followed by a possessive quantifier.
1791
1792 15. Modified the Makefile to add libpcre.la as a prerequisite for
1793 libpcreposix.la because I was told this is needed for a parallel build to
1794 work.
1795
1796 16. If a pattern that contained .* following optional items at the start was
1797 studied, the wrong optimizing data was generated, leading to matching
1798 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1799 matching string must start with a or b or c. The correct conclusion for
1800 this pattern is that a match can start with any character.
1801
1802
1803 Version 4.4 13-Aug-03
1804 ---------------------
1805
1806 1. In UTF-8 mode, a character class containing characters with values between
1807 127 and 255 was not handled correctly if the compiled pattern was studied.
1808 In fixing this, I have also improved the studying algorithm for such
1809 classes (slightly).
1810
1811 2. Three internal functions had redundant arguments passed to them. Removal
1812 might give a very teeny performance improvement.
1813
1814 3. Documentation bug: the value of the capture_top field in a callout is *one
1815 more than* the number of the hightest numbered captured substring.
1816
1817 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1818 in incorrectly linking with a previously installed version. They now link
1819 explicitly with libpcre.la.
1820
1821 5. configure.in no longer needs to recognize Cygwin specially.
1822
1823 6. A problem in pcre.in for Windows platforms is fixed.
1824
1825 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1826 pcretest, it used to include the size of the study block as part of its
1827 output. Unfortunately, the structure contains a field that has a different
1828 size on different hardware architectures. This meant that the tests that
1829 showed this size failed. As the block is currently always of a fixed size,
1830 this information isn't actually particularly useful in pcretest output, so
1831 I have just removed it.
1832
1833 8. Three pre-processor statements accidentally did not start in column 1.
1834 Sadly, there are *still* compilers around that complain, even though
1835 standard C has not required this for well over a decade. Sigh.
1836
1837 9. In pcretest, the code for checking callouts passed small integers in the
1838 callout_data field, which is a void * field. However, some picky compilers
1839 complained about the casts involved for this on 64-bit systems. Now
1840 pcretest passes the address of the small integer instead, which should get
1841 rid of the warnings.
1842
1843 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1844 both compile and run time, and gives an error if an invalid UTF-8 sequence
1845 is found. There is a option for disabling this check in cases where the
1846 string is known to be correct and/or the maximum performance is wanted.
1847
1848 11. In response to a bug report, I changed one line in Makefile.in from
1849
1850 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1851 to
1852 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1853
1854 to look similar to other lines, but I have no way of telling whether this
1855 is the right thing to do, as I do not use Windows. No doubt I'll get told
1856 if it's wrong...
1857
1858
1859 Version 4.3 21-May-03
1860 ---------------------
1861
1862 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1863 Makefile.
1864
1865 2. Some refactoring to improve the quality of the code:
1866
1867 (i) The utf8_table... variables are now declared "const".
1868
1869 (ii) The code for \cx, which used the "case flipping" table to upper case
1870 lower case letters, now just substracts 32. This is ASCII-specific,
1871 but the whole concept of \cx is ASCII-specific, so it seems
1872 reasonable.
1873
1874 (iii) PCRE was using its character types table to recognize decimal and
1875 hexadecimal digits in the pattern. This is silly, because it handles
1876 only 0-9, a-f, and A-F, but the character types table is locale-
1877 specific, which means strange things might happen. A private
1878 table is now used for this - though it costs 256 bytes, a table is
1879 much faster than multiple explicit tests. Of course, the standard
1880 character types table is still used for matching digits in subject
1881 strings against \d.
1882
1883 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1884 ending in _t are). So I've renamed it as ESC_tee.
1885
1886 3. The first argument for regexec() in the POSIX wrapper should have been
1887 defined as "const".
1888
1889 4. Changed pcretest to use malloc() for its buffers so that they can be
1890 Electric Fenced for debugging.
1891
1892 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1893 to read one or more bytes before the start of the subject string. Often this
1894 had no effect on PCRE's behaviour, but in some circumstances it could
1895 provoke a segmentation fault.
1896
1897 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1898 to try to read one or more bytes before the start of the subject string.
1899
1900 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1901 UTF-8 support could misbehave in various ways if the subject string
1902 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1903 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1904 back over UTF-8 characters.)
1905
1906
1907 Version 4.2 14-Apr-03
1908 ---------------------
1909
1910 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1911
1912 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1913 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1914 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1915 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1916 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1917 and BUILD_EXEEXT
1918 Note: automatic setting of the BUILD variables is not yet working
1919 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1920 compile-time but not at link-time
1921 [LINK]: use for linking executables only
1922 make different versions for Windows and non-Windows
1923 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1924 libraries
1925 [LINK_FOR_BUILD]: new variable
1926 [OBJEXT]: use throughout
1927 [EXEEXT]: use throughout
1928 <winshared>: new target
1929 <wininstall>: new target
1930 <dftables.o>: use native compiler
1931 <dftables>: use native linker
1932 <install>: handle Windows platform correctly
1933 <clean>: ditto
1934 <check>: ditto
1935 copy DLL to top builddir before testing
1936
1937 As part of these changes, -no-undefined was removed again. This was reported
1938 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1939 in any case.
1940
1941 3. Some tidies to get rid of compiler warnings:
1942
1943 . In the match_data structure, match_limit was an unsigned long int, whereas
1944 match_call_count was an int. I've made them both unsigned long ints.
1945
1946 . In pcretest the fact that a const uschar * doesn't automatically cast to
1947 a void * provoked a warning.
1948
1949 . Turning on some more compiler warnings threw up some "shadow" variables
1950 and a few more missing casts.
1951
1952 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1953 option, a class that contained a single character with a value between 128
1954 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1955
1956 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1957 option, a class that contained several characters, but with at least one
1958 whose value was between 128 and 255 caused PCRE to crash.
1959
1960
1961 Version 4.1 12-Mar-03
1962 ---------------------
1963
1964 1. Compiling with gcc -pedantic found a couple of places where casts were
1965 needed, and a string in dftables.c that was longer than standard compilers are
1966 required to support.
1967
1968 2. Compiling with Sun's compiler found a few more places where the code could
1969 be tidied up in order to avoid warnings.
1970
1971 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1972 first of these names is deprecated in the latest Autoconf in favour of the name
1973 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1974 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1975 analogy I have changed it to CFLAGS_FOR_BUILD.
1976
1977 4. Added -no-undefined to the linking command in the Makefile, because this is
1978 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1979 linking step for the pcreposix library.
1980
1981 5. PCRE was failing to diagnose the case of two named groups with the same
1982 name.
1983
1984 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1985 literal character that is needed in the subject for a match, and scans along to
1986 ensure that it is present before embarking on the full matching process. This
1987 saves time in cases of nested unlimited repeats that are never going to match.
1988 Problem: the scan can take a lot of time if the subject is very long (e.g.
1989 megabytes), thus penalizing straightforward matches. It is now done only if the
1990 amount of subject to be scanned is less than 1000 bytes.
1991
1992 7. A lesser problem with the same optimization is that it was recording the
1993 first character of an anchored pattern as "needed", thus provoking a search
1994 right along the subject, even when the first match of the pattern was going to
1995 fail. The "needed" character is now not set for anchored patterns, unless it
1996 follows something in the pattern that is of non-fixed length. Thus, it still
1997 fulfils its original purpose of finding quick non-matches in cases of nested
1998 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1999
2000
2001 Version 4.0 17-Feb-03
2002 ---------------------
2003
2004 1. If a comment in an extended regex that started immediately after a meta-item
2005 extended to the end of string, PCRE compiled incorrect data. This could lead to
2006 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2007
2008 2. Moved to autoconf 2.53 and libtool 1.4.2.
2009
2010 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2011 the special perltest8 script is no longer needed - all the tests can be run
2012 from a single perltest script.
2013
2014 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2015 by \s. It has now been removed in PCRE. This means it isn't recognized as
2016 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2017 class [:space:] *does* include VT, thereby creating a mess.
2018
2019 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2020 space and tab.
2021
2022 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2023 its new features into the main test script, reducing the number of scripts.
2024
2025 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2026 were backward compatible, and made the (?i) apply to the whole pattern, as if
2027 /i were given. Now it behaves more logically, and applies the option setting
2028 only to what follows. PCRE has been changed to follow suit. However, if it
2029 finds options settings right at the start of the pattern, it extracts them into
2030 the global options, as before. Thus, they show up in the info data.
2031
2032 8. Added support for the \Q...\E escape sequence. Characters in between are
2033 treated as literals. This is slightly different from Perl in that $ and @ are
2034 also handled as literals inside the quotes. In Perl, they will cause variable
2035 interpolation. Note the following examples:
2036
2037 Pattern PCRE matches Perl matches
2038
2039 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2040 \Qabc\$xyz\E abc\$xyz abc\$xyz
2041 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2042
2043 For compatibility with Perl, \Q...\E sequences are recognized inside character
2044 classes as well as outside them.
2045
2046 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2047 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2048 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2049 signed/unsigned warnings.
2050
2051 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2052 option for pcretest, so I've replaced it by a simple function that does just
2053 that job.
2054
2055 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2056 "pcregrep -".
2057
2058 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2059 Java package. This provides some syntactic sugar for simple cases of what my
2060 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2061 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2062 item, you can use this simplified notation. Note that only makes sense with
2063 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2064 greediness, whatever the setting of the PCRE_UNGREEDY option.
2065
2066 13. A change of greediness default within a pattern was not taking effect at
2067 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2068 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2069 was abstracted outside.
2070
2071 14. PCRE now supports the \G assertion. It is true when the current matching
2072 position is at the start point of the match. This differs from \A when the
2073 starting offset is non-zero. Used with the /g option of pcretest (or similar
2074 code), it works in the same way as it does for Perl's /g option. If all
2075 alternatives of a regex begin with \G, the expression is anchored to the start
2076 match position, and the "anchored" flag is set in the compiled expression.
2077
2078 15. Some bugs concerning the handling of certain option changes within patterns
2079 have been fixed. These applied to options other than (?ims). For example,
2080 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2081 been the other way round. Some of this was related to change 7 above.
2082
2083 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2084 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2085 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2086 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2087
2088 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2089 mode. Unlike ".", it always matches newline, whatever the setting of
2090 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2091 assertions. Perl allows it, but it doesn't (in general) work because it can't
2092 calculate the length of the lookbehind. At least, that's the case for Perl
2093 5.8.0 - I've been told they are going to document that it doesn't work in
2094 future.
2095
2096 18. Added an error diagnosis for escapes that PCRE does not support: these are
2097 \L, \l, \N, \P, \p, \U, \u, and \X.
2098
2099 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2100 reading past the end of the pattern in cases such as /[abcd/.
2101
2102 20. PCRE was getting more memory than necessary for patterns with classes that
2103 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2104
2105 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2106 compiling PCRE for use with Virtual Pascal.
2107
2108 22. Small fix to the Makefile to make it work properly if the build is done
2109 outside the source tree.
2110
2111 23. Added a new extension: a condition to go with recursion. If a conditional
2112 subpattern starts with (?(R) the "true" branch is used if recursion has
2113 happened, whereas the "false" branch is used only at the top level.
2114
2115 24. When there was a very long string of literal characters (over 255 bytes
2116 without UTF support, over 250 bytes with UTF support), the computation of how
2117 much memory was required could be incorrect, leading to segfaults or other
2118 strange effects.
2119
2120 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2121 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2122 there was a subsequent back reference to those brackets. This meant that, for
2123 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2124 possible to check for precisely this case. All we can do is abandon the
2125 optimization if .* occurs inside capturing brackets when there are any back
2126 references whatsoever. (See below for a better fix that came later.)
2127
2128 26. The handling of the optimization for finding the first character of a
2129 non-anchored pattern, and for finding a character that is required later in the
2130 match were failing in some cases. This didn't break the matching; it just
2131 failed to optimize when it could. The way this is done has been re-implemented.
2132
2133 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2134
2135 28. Added a new feature that provides some of the functionality that Perl
2136 provides with (?{...}). The facility is termed a "callout". The way it is done
2137 in PCRE is for the caller to provide an optional function, by setting
2138 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2139 global variable. By default it is unset, which disables all calling out. To get
2140 the function called, the regex must include (?C) at appropriate points. This
2141 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2142 This provides a means of identifying different callout points. When PCRE
2143 reaches such a point in the regex, if pcre_callout has been set, the external
2144 function is called. It is provided with data in a structure called
2145 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2146 matching continues; if it returns a non-zero value, the match at the current
2147 point fails. However, backtracking will occur if possible. [This was changed
2148 later and other features added - see item 49 below.]
2149
2150 29. pcretest is upgraded to test the callout functionality. It provides a
2151 callout function that displays information. By default, it shows the start of
2152 the match and the current position in the text. There are some new data escapes
2153 to vary what happens:
2154
2155 \C+ in addition, show current contents of captured substrings
2156 \C- do not supply a callout function
2157 \C!n return 1 when callout number n is reached
2158 \C!n!m return 1 when callout number n is reached for the mth time
2159
2160 30. If pcregrep was called with the -l option and just a single file name, it
2161 output "<stdin>" if a match was found, instead of the file name.
2162
2163 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2164 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2165 pcre_exec(). This saves a malloc/free per call. The default value of
2166 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2167 when configuring.
2168
2169 32. The default maximum size of a compiled pattern is 64K. There have been a
2170 few cases of people hitting this limit. The code now uses macros to handle the
2171 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2172 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2173 configuring. Tests 2 and 5 work only with 2-byte links because they output
2174 debugging information about compiled patterns.
2175
2176 33. Internal code re-arrangements:
2177
2178 (a) Moved the debugging function for printing out a compiled regex into
2179 its own source file (printint.c) and used #include to pull it into
2180 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2181 separate copies.
2182
2183 (b) Defined the list of op-code names for debugging as a macro in
2184 internal.h so that it is next to the definition of the opcodes.
2185
2186 (c) Defined a table of op-code lengths for simpler skipping along compiled
2187 code. This is again a macro in internal.h so that it is next to the
2188 definition of the opcodes.
2189
2190 34. Added support for recursive calls to individual subpatterns, along the
2191 lines of Robin Houston's patch (but implemented somewhat differently).
2192
2193 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2194 allow it to read and process whole directories in Win32. This code was
2195 contributed by Lionel Fourquaux; it has not been tested by me.
2196
2197 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2198 used to name a group. Names consist of alphanumerics and underscores, and must
2199 be unique. Back references use the syntax (?P=name) and recursive calls use
2200 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2201 numbers. The function pcre_fullinfo() can be used after compilation to extract
2202 a name/number map. There are three relevant calls:
2203
2204 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2205 PCRE_INFO_NAMECOUNT yields the number of entries
2206 PCRE_INFO_NAMETABLE yields a pointer to the map.
2207
2208 The map is a vector of fixed-size entries. The size of each entry depends on
2209 the length of the longest name used. The first two bytes of each entry are the
2210 group number, most significant byte first. There follows the corresponding
2211 name, zero terminated. The names are in alphabetical order.
2212
2213 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2214 case instead of 255. Making it the same both with and without UTF-8 support
2215 means that the same test output works with both.
2216
2217 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2218 calling malloc() with a zero argument.
2219
2220 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2221 optimization. I've improved things by keeping a bitmap of backreferences with
2222 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2223 fact referenced, the optimization can be applied. It is unlikely that a
2224 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2225 the match to follow \n) will appear inside brackets with a number greater than
2226 31, but if it does, any back reference > 31 suppresses the optimization.
2227
2228 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2229 of disabling numbered capturing parentheses. Any opening parenthesis that is
2230 not followed by ? behaves as if it were followed by ?: but named parentheses
2231 can still be used for capturing (and they will acquire numbers in the usual
2232 way).
2233
2234 41. Redesigned the return codes from the match() function into yes/no/error so
2235 that errors can be passed back from deep inside the nested calls. A malloc
2236 failure while inside a recursive subpattern call now causes the
2237 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2238
2239 42. It is now possible to set a limit on the number of times the match()
2240 function is called in a call to pcre_exec(). This facility makes it possible to
2241 limit the amount of recursion and backtracking, though not in a directly
2242 obvious way, because the match() function is used in a number of different
2243 circumstances. The count starts from zero for each position in the subject
2244 string (for non-anchored patterns). The default limit is, for compatibility, a
2245 large number, namely 10 000 000. You can change this in two ways:
2246
2247 (a) When configuring PCRE before making, you can use --with-match-limit=n
2248 to set a default value for the compiled library.
2249
2250 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2251 a different value is set. See 45 below.
2252
2253 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2254
2255 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2256 of things that can be changed at compile time. The first argument specifies
2257 what is wanted and the second points to where the information is to be placed.
2258 The current list of available information is:
2259
2260 PCRE_CONFIG_UTF8
2261
2262 The output is an integer that is set to one if UTF-8 support is available;
2263 otherwise it is set to zero.
2264
2265 PCRE_CONFIG_NEWLINE
2266
2267 The output is an integer that it set to the value of the code that is used for
2268 newline. It is either LF (10) or CR (13).
2269
2270 PCRE_CONFIG_LINK_SIZE
2271
2272 The output is an integer that contains the number of bytes used for internal
2273 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2274
2275 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2276
2277 The output is an integer that contains the threshold above which the POSIX
2278 interface uses malloc() for output vectors. See item 31 above.
2279
2280 PCRE_CONFIG_MATCH_LIMIT
2281
2282 The output is an unsigned integer that contains the default limit of the number
2283 of match() calls in a pcre_exec() execution. See 42 above.
2284
2285 44. pcretest has been upgraded by the addition of the -C option. This causes it
2286 to extract all the available output from the new pcre_config() function, and to
2287 output it. The program then exits immediately.
2288
2289 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2290 order to support additional features. One way would have been to define
2291 pcre_exec2() (for example) with extra arguments, but this would not have been
2292 extensible, and would also have required all calls to the original function to
2293 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2294 is used for passing in "extra" data from pcre_study().
2295
2296 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2297 contains the following fields:
2298
2299 flags a bitmap indicating which of the following fields are set
2300 study_data opaque data from pcre_study()
2301 match_limit a way of specifying a limit on match() calls for a specific
2302 call to pcre_exec()
2303 callout_data data for callouts (see 49 below)
2304
2305 The flag bits are also defined in pcre.h, and are
2306
2307 PCRE_EXTRA_STUDY_DATA
2308 PCRE_EXTRA_MATCH_LIMIT
2309 PCRE_EXTRA_CALLOUT_DATA
2310
2311 The pcre_study() function now returns one of these new pcre_extra blocks, with
2312 the actual study data pointed to by the study_data field, and the
2313 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2314 before. That is, this change is entirely upwards-compatible and requires no
2315 change to existing code.
2316
2317 If you want to pass in additional data to pcre_exec(), you can either place it
2318 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2319 block.
2320
2321 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2322 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2323 times with different match limits, until it finds the minimum value needed for
2324 pcre_exec() to complete. The value is then output. This can be instructive; for
2325 most simple matches the number is quite small, but for pathological cases it
2326 gets very large very quickly.
2327
2328 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2329 returns the size of the data block pointed to by the study_data field in a
2330 pcre_extra block, that is, the value that was passed as the argument to
2331 pcre_malloc() when PCRE was getting memory in which to place the information
2332 created by pcre_study(). The fourth argument should point to a size_t variable.
2333 pcretest has been extended so that this information is shown after a successful
2334 pcre_study() call when information about the compiled regex is being displayed.
2335
2336 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2337 because what follows is always an absolute path. (Later: it turns out that this
2338 is more than cosmetic for MinGW, because it doesn't like empty path
2339 components.)
2340
2341 49. Some changes have been made to the callout feature (see 28 above):
2342
2343 (i) A callout function now has three choices for what it returns:
2344
2345 0 => success, carry on matching
2346 > 0 => failure at this point, but backtrack if possible
2347 < 0 => serious error, return this value from pcre_exec()
2348
2349 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2350 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2351 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2352 use by callout functions. It will never be used by PCRE itself.
2353
2354 (ii) The pcre_extra structure (see 45 above) has a void * field called
2355 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2356 pcre_callout_block structure has a field of the same name. The contents of
2357 the field passed in the pcre_extra structure are passed to the callout
2358 function in the corresponding field in the callout block. This makes it
2359 easier to use the same callout-containing regex from multiple threads. For
2360 testing, the pcretest program has a new data escape
2361
2362 \C*n pass the number n (may be negative) as callout_data
2363
2364 If the callout function in pcretest receives a non-zero value as
2365 callout_data, it returns that value.
2366
2367 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2368 there were some redundant $(CFLAGS) in commands that are now specified as
2369 $(LINK), which already includes $(CFLAGS).
2370
2371 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2372 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2373 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2374 one-byte characters throughout. Note that case-insensitive matching applies
2375 only to characters whose values are less than 256. PCRE doesn't support the
2376 notion of cases for higher-valued characters.
2377
2378 (i) A character class whose characters are all within 0-255 is handled as
2379 a bit map, and the map is inverted for negative classes. Previously, a
2380 character > 255 always failed to match such a class; however it should
2381 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2382
2383 (ii) A negated character class with a single character < 255 is coded as
2384 "not this character" (OP_NOT). This wasn't working properly when the test
2385 character was multibyte, either singly or repeated.
2386
2387 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2388 mode, for example: \x{100}{2,3}.
2389
2390 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2391 singly or repeated) now correctly test multibyte characters. However,
2392 PCRE doesn't recognize any characters with values greater than 255 as
2393 digits, spaces, or word characters. Such characters always match \D, \S,
2394 and \W, and never match \d, \s, or \w.
2395
2396 (v) Classes may now contain characters and character ranges with values
2397 greater than 255. For example: [ab\x{100}-\x{400}].
2398
2399 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2400 PCRE in UTF-8 mode.
2401
2402 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2403 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2404 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2405 value.)
2406
2407 53. The single man page has become too large. I have therefore split it up into
2408 a number of separate man pages. These also give rise to individual HTML pages;
2409 these are now put in a separate directory, and there is an index.html page that
2410 lists them all. Some hyperlinking between the pages has been installed.
2411
2412 54. Added convenience functions for handling named capturing parentheses.
2413
2414 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2415 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2416 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2417 are faulted.
2418
2419 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2420 calling configure. These values are used when compiling the dftables.c program
2421 which is run to generate the source of the default character tables. They
2422 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2423 you will need to set these values.
2424
2425 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2426
2427
2428 Version 3.9 02-Jan-02
2429 ---------------------
2430
2431 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2432
2433 2. If --disable-static was given, the building process failed when trying to
2434 build pcretest and pcregrep. (For some reason it was using libtool to compile
2435 them, which is not right, as they aren't part of the library.)
2436
2437
2438 Version 3.8 18-Dec-01
2439 ---------------------
2440
2441 1. The experimental UTF-8 code was completely screwed up. It was packing the
2442 bytes in the wrong order. How dumb can you get?
2443
2444
2445 Version 3.7 29-Oct-01
2446 ---------------------
2447
2448 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2449 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2450 this didn't happen under Solaris 8, where I normally test things.
2451
2452 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2453 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2454 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2455 this fix an hour or so after the initial 3.7 release.)
2456
2457
2458 Version 3.6 23-Oct-01
2459 ---------------------
2460
2461 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2462 offsets passed as NULL with zero offset count.
2463
2464 2. The config.guess and config.sub files had not been updated when I moved to
2465 the latest autoconf.
2466
2467
2468 Version 3.5 15-Aug-01
2469 ---------------------
2470
2471 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2472 had been forgotten.
2473
2474 2. By using declared but undefined structures, we can avoid using "void"
2475 definitions in pcre.h while keeping the internal definitions of the structures
2476 private.
2477
2478 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2479 user point of view, this means that both static and shared libraries are built
2480 by default, but this can be individually controlled. More of the work of
2481 handling this static/shared cases is now inside libtool instead of PCRE's make
2482 file.
2483
2484 4. The pcretest utility is now installed along with pcregrep because it is
2485 useful for users (to test regexs) and by doing this, it automatically gets
2486 relinked by libtool. The documentation has been turned into a man page, so
2487 there are now .1, .txt, and .html versions in /doc.
2488
2489 5. Upgrades to pcregrep:
2490 (i) Added long-form option names like gnu grep.
2491 (ii) Added --help to list all options with an explanatory phrase.
2492 (iii) Added -r, --recursive to recurse into sub-directories.
2493 (iv) Added -f, --file to read patterns from a file.
2494
2495 6. pcre_exec() was referring to its "code" argument before testing that
2496 argument for NULL (and giving an error if it was NULL).
2497
2498 7. Upgraded Makefile.in to allow for compiling in a different directory from
2499 the source directory.
2500
2501 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2502 options bits, the pointer it was passed was to an int instead of to an unsigned
2503 long int. This mattered only on 64-bit systems.
2504
2505 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2506 generated) instead of pcre.in, which it its source. Also made the same change
2507 in several of the .c files.
2508
2509 10. A new release of gcc defines printf() as a macro, which broke pcretest
2510 because it had an ifdef in the middle of a string argument for printf(). Fixed
2511 by using separate calls to printf().
2512
2513 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2514 script, to force use of CR or LF instead of \n in the source. On non-Unix
2515 systems, the value can be set in config.h.
2516
2517 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2518 absolute limit. Changed the text of the error message to make this clear, and
2519 likewise updated the man page.
2520
2521 13. The limit of 99 on the number of capturing subpatterns has been removed.
2522 The new limit is 65535, which I hope will not be a "real" limit.
2523
2524
2525 Version 3.4 22-Aug-00
2526 ---------------------
2527
2528 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2529
2530 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2531
2532
2533 Version 3.3 01-Aug-00
2534 ---------------------
2535
2536 1. If an octal character was given, but the value was greater than \377, it
2537 was not getting masked to the least significant bits, as documented. This could
2538 lead to crashes in some systems.
2539
2540 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2541 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2542
2543 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2544 These just pass their arguments on to (pcre_free)(), but they are provided
2545 because some uses of PCRE bind it to non-C systems that can call its functions,
2546 but cannot call free() or pcre_free() directly.
2547
2548 4. Add "make test" as a synonym for "make check". Corrected some comments in
2549 the Makefile.
2550
2551 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2552 Makefile.
2553
2554 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2555 command called pgrep for grepping around the active processes.
2556
2557 7. Added the beginnings of support for UTF-8 character strings.
2558
2559 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2560 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2561 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2562 out for the ar command.)
2563
2564
2565 Version 3.2 12-May-00
2566 ---------------------
2567
2568 This is purely a bug fixing release.
2569
2570 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2571 of ZA. This was just one example of several cases that could provoke this bug,
2572 which was introduced by change 9 of version 2.00. The code for breaking
2573 infinite loops after an iteration that matches an empty string was't working
2574 correctly.
2575
2576 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2577 when matched against abbab (for example). After matching an empty string, it
2578 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2579 caused it to match further down the string than it should.
2580
2581 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2582 was there because it doesn't seem to be needed, and it causes trouble on some
2583 systems, as it is not a Standard C header. It has been removed.
2584
2585 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2586 were reported on the Macintosh. The changes were from
2587
2588 while ((c = *(++ptr)) != 0 && c != '\n');
2589 to
2590 while ((c = *(++ptr)) != 0 && c != '\n') ;
2591
2592 Totally extraordinary, but if that's what it takes...
2593
2594 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2595 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2596 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2597 assumes the way PCRE uses memmove() (always moving upwards).
2598
2599 6. PCRE is being used in one environment where strchr() is not available. There
2600 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2601 faster code anyway.
2602
2603
2604 Version 3.1 09-Feb-00
2605 ---------------------
2606
2607 The only change in this release is the fixing of some bugs in Makefile.in for
2608 the "install" target:
2609
2610 (1) It was failing to install pcreposix.h.
2611
2612 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2613
2614
2615 Version 3.0 01-Feb-00
2616 ---------------------
2617
2618 1. Add support for the /+ modifier to perltest (to output $` like it does in
2619 pcretest).
2620
2621 2. Add support for the /g modifier to perltest.
2622
2623 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2624 matches null strings.
2625
2626 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2627 pattern. Perl treats empty patterns specially - it reuses the most recent
2628 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2629 effect.
2630
2631 5. The POSIX interface was broken in that it was just handing over the POSIX
2632 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2633 required a bigger vector, with some working space on the end. This means that
2634 the POSIX wrapper now has to get and free some memory, and copy the results.
2635
2636 6. Added some simple autoconf support, placing the test data and the
2637 documentation in separate directories, re-organizing some of the
2638 information files, and making it build pcre-config (a GNU standard). Also added
2639 libtool support for building PCRE as a shared library, which is now the
2640 default.
2641
2642 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2643 09 are not valid octal constants. Single digits will be used for minor values
2644 less than 10.
2645
2646 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2647 existing programs that set these in the POSIX interface can use PCRE without
2648 modification.
2649
2650 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2651 return all that pcre_info() returns, plus additional data. The pcre_info()
2652 function is retained for compatibility, but is considered to be obsolete.
2653
2654 10. Added experimental recursion feature (?R) to handle one common case that
2655 Perl 5.6 will be able to do with (?p{...}).
2656
2657 11. Added support for POSIX character classes like [:alpha:], which Perl is
2658 adopting.
2659
2660
2661 Version 2.08 31-Aug-99
2662 ----------------------
2663
2664 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2665 trying to match at the startoffset position, but instead was moving forward to
2666 the next newline as if a previous match had failed.
2667
2668 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2669 and could get into a loop if a null string was matched other than at the start
2670 of the subject.
2671
2672 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2673 be distinguished at compile time, and for completeness also added PCRE_DATE.
2674
2675 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2676 in GnuWin32 environments.
2677
2678
2679 Version 2.07 29-Jul-99
2680 ----------------------
2681
2682 1. The documentation is now supplied in plain text form and HTML as well as in
2683 the form of man page sources.
2684
2685 2. C++ compilers don't like assigning (void *) values to other pointer types.
2686 In particular this affects malloc(). Although there is no problem in Standard
2687 C, I've put in casts to keep C++ compilers happy.
2688
2689 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2690 should be (const char *).
2691
2692 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2693 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2694 However, I haven't made this a standard facility. The documentation doesn't
2695 mention it, and the Makefile doesn't support it.
2696
2697 5. The Makefile now contains an "install" target, with editable destinations at
2698 the top of the file. The pcretest program is not installed.
2699
2700 6. pgrep -V now gives the PCRE version number and date.
2701
2702 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2703 causing the entire string to be ignored, instead of just the last character.
2704
2705 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2706 non-matching string, it can take a very, very long time, even for strings of
2707 quite modest length, because of the nested recursion. PCRE now does better in
2708 some of these cases. It does this by remembering the last required literal
2709 character in the pattern, and pre-searching the subject to ensure it is present
2710 before running the real match. In other words, it applies a heuristic to detect
2711 some types of certain failure quickly, and in the above example, if presented
2712 with a string that has no trailing " it gives "no match" very quickly.
2713
2714 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2715 other alternatives are tried instead.
2716
2717
2718 Version 2.06 09-Jun-99
2719 ----------------------
2720
2721 1. Change pcretest's output for amount of store used to show just the code
2722 space, because the remainder (the data block) varies in size between 32-bit and
2723 64-bit systems.
2724
2725 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2726 start matching at. This allows lookbehinds to work when searching for multiple
2727 occurrences in a string.
2728
2729 3. Added additional options to pcretest for testing multiple occurrences:
2730
2731 /+ outputs the rest of the string that follows a match
2732 /g loops for multiple occurrences, using the new startoffset argument
2733 /G loops for multiple occurrences by passing an incremented pointer
2734
2735 4. PCRE wasn't doing the "first character" optimization for patterns starting
2736 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2737 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2738 the letter 'x'. On long subject strings, this gives a significant speed-up.
2739
2740
2741 Version 2.05 21-Apr-99
2742 ----------------------
2743
2744 1. Changed the type of magic_number from int to long int so that it works
2745 properly on 16-bit systems.
2746
2747 2. Fixed a bug which caused patterns starting with .* not to work correctly
2748 when the subject string contained newline characters. PCRE was assuming
2749 anchoring for such patterns in all cases, which is not correct because .* will
2750 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2751 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2752 must be retried after every newline in the subject.
2753
2754
2755 Version 2.04 18-Feb-99
2756 ----------------------
2757
2758 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2759 computation of the store needed to hold the pattern was incorrect (too large).
2760 If such patterns were nested a few deep, this could multiply and become a real
2761 problem.
2762
2763 2. Added /M option to pcretest to show the memory requirement of a specific
2764 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2765
2766 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2767 compiled in such a way that the backtracking after subsequent failure was
2768 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2769 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2770
2771
2772 Version 2.03 02-Feb-99
2773 ----------------------
2774
2775 1. Fixed typo and small mistake in man page.
2776
2777 2. Added 4th condition (GPL supersedes if conflict) and created separate
2778 LICENCE file containing the conditions.
2779
2780 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2781 Perl, that is the internal \ allows the delimiter to be included in the
2782 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2783 the final delimiter, add \ to the end of the pattern (to test the error).
2784
2785 4. Added the convenience functions for extracting substrings after a successful
2786 match. Updated pcretest to make it able to test these functions.
2787
2788
2789 Version 2.02 14-Jan-99
2790 ----------------------
2791
2792 1. Initialized the working variables associated with each extraction so that
2793 their saving and restoring doesn't refer to uninitialized store.
2794
2795 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2796 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2797 fix the problem.
2798
2799 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2800 calls, and wasn't printing the correct value for compiling calls. Increased the
2801 default value of LOOPREPEAT, and the number of significant figures in the
2802 times.
2803
2804 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2805
2806 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2807 a building problem on Windows NT with a FAT file system.
2808
2809
2810 Version 2.01 21-Oct-98
2811 ----------------------
2812
2813 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2814 to character tables built by pcre_maketables() in the current locale. If NULL
2815 is passed, the default tables are used.
2816
2817
2818 Version 2.00 24-Sep-98
2819 ----------------------
2820
2821 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2822 it any more.
2823
2824 2. Allow quantification of (?>) groups, and make it work correctly.
2825
2826 3. The first character computation wasn't working for (?>) groups.
2827
2828 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2829 end of the subject) and add 5.005's \z, which really does match only at the
2830 very end of the subject.
2831
2832 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2833
2834 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2835 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2836 localized options. All options to pcre_study() were also removed.
2837
2838 7. Add other new features from 5.005:
2839
2840 $(?<= positive lookbehind
2841 $(?<! negative lookbehind
2842 (?imsx-imsx) added the unsetting capability
2843 such a setting is global if at outer level; local otherwise
2844 (?imsx-imsx:) non-capturing groups with option setting
2845 (?(cond)re|re) conditional pattern matching
2846
2847 A backreference to itself in a repeated group matches the previous
2848 captured string.
2849
2850 8. General tidying up of studying (both automatic and via "study")
2851 consequential on the addition of new assertions.
2852
2853 9. As in 5.005, unlimited repeated groups that could match an empty substring
2854 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2855 runtime if any iteration does actually match an empty substring.
2856
2857 10. Include the RunTest script in the distribution.
2858
2859 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2860 discrepancies, some of which were old and were also with respect to 5.004. They
2861 have now been fixed.
2862
2863
2864 Version 1.09 28-Apr-98
2865 ----------------------
2866
2867 1. A negated single character class followed by a quantifier with a minimum
2868 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2869 program crashes, or just wrong answers. This did not apply to negated classes
2870 containing more than one character, or to minima other than one.
2871
2872
2873 Version 1.08 27-Mar-98
2874 ----------------------
2875
2876 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2877
2878 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2879 latter must appear before anything that relies on it in the pattern.
2880
2881
2882 Version 1.07 16-Feb-98
2883 ----------------------
2884
2885 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2886 repeat of a potentially empty string).
2887
2888
2889 Version 1.06 23-Jan-98
2890 ----------------------
2891
2892 1. Added Markus Oberhumer's little patches for C++.
2893
2894 2. Literal strings longer than 255 characters were broken.
2895
2896
2897 Version 1.05 23-Dec-97
2898 ----------------------
2899
2900 1. Negated character classes containing more than one character were failing if
2901 PCRE_CASELESS was set at run time.
2902
2903
2904 Version 1.04 19-Dec-97
2905 ----------------------
2906
2907 1. Corrected the man page, where some "const" qualifiers had been omitted.
2908
2909 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2910 input syntax.
2911
2912 3. Fixed memory leak which occurred when a regex with back references was
2913 matched with an offsets vector that wasn't big enough. The temporary memory
2914 that is used in this case wasn't being freed if the match failed.
2915
2916 4. Tidied pcretest to ensure it frees memory that it gets.
2917
2918 5. Temporary memory was being obtained in the case where the passed offsets
2919 vector was exactly big enough.
2920
2921 6. Corrected definition of offsetof() from change 5 below.
2922
2923 7. I had screwed up change 6 below and broken the rules for the use of
2924 setjmp(). Now fixed.
2925
2926
2927 Version 1.03 18-Dec-97
2928 ----------------------
2929
2930 1. A erroneous regex with a missing opening parenthesis was correctly
2931 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2932 on some systems.
2933
2934 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2935 it was reported that one broken compiler failed on the former because "code" is
2936 also an independent variable.
2937
2938 3. The erroneous regex a[]b caused an array overrun reference.
2939
2940 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2941 fail on data ending with that character. (It was going on too far, and checking
2942 the next character, typically a binary zero.) This was specific to the
2943 optimized code for single-character negative classes.
2944
2945 5. Added a contributed patch from the TIN world which does the following:
2946
2947 + Add an undef for memmove, in case the the system defines a macro for it.
2948
2949 + Add a definition of offsetof(), in case there isn't one. (I don't know
2950 the reason behind this - offsetof() is part of the ANSI standard - but
2951 it does no harm).
2952
2953 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2954 most of the places where whitespace preceded '#'. I have given up and
2955 allowed the remaining 2 cases to be at the margin.
2956
2957 + Rename some variables in pcre to eliminate shadowing. This seems very
2958 pedantic, but does no harm, of course.
2959
2960 6. Moved the call to setjmp() into its own function, to get rid of warnings
2961 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2962
2963 7. Constructs such as \d{8,} were compiling into the equivalent of
2964 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2965 outcome, but in this particular case used more store than had been allocated,
2966 which caused the bug to be discovered because it threw up an internal error.
2967
2968 8. The debugging code in both pcre and pcretest for outputting the compiled
2969 form of a regex was going wrong in the case of back references followed by
2970 curly-bracketed repeats.
2971
2972
2973 Version 1.02 12-Dec-97
2974 ----------------------
2975
2976 1. Typos in pcre.3 and comments in the source fixed.
2977
2978 2. Applied a contributed patch to get rid of places where it used to remove
2979 'const' from variables, and fixed some signed/unsigned and uninitialized
2980 variable warnings.
2981
2982 3. Added the "runtest" target to Makefile.
2983
2984 4. Set default compiler flag to -O2 rather than just -O.
2985
2986
2987 Version 1.01 19-Nov-97
2988 ----------------------
2989
2990 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2991 like /([ab]*)*/, that is, for classes with more than one character in them.
2992
2993 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2994 as /((?>a*))*/ (a PCRE_EXTRA facility).
2995
2996
2997 Version 1.00 18-Nov-97
2998 ----------------------
2999
3000 1. Added compile-time macros to support systems such as SunOS4 which don't have
3001 memmove() or strerror() but have other things that can be used instead.
3002
3003 2. Arranged that "make clean" removes the executables.
3004
3005
3006 Version 0.99 27-Oct-97
3007 ----------------------
3008
3009 1. Fixed bug in code for optimizing classes with only one character. It was
3010 initializing a 32-byte map regardless, which could cause it to run off the end
3011 of the memory it had got.
3012
3013 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3014
3015
3016 Version 0.98 22-Oct-97
3017 ----------------------
3018
3019 1. Fixed bug in code for handling temporary memory usage when there are more
3020 back references than supplied space in the ovector. This could cause segfaults.
3021
3022
3023 Version 0.97 21-Oct-97
3024 ----------------------
3025
3026 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3027
3028 2. Optimized negated single characters not to use a bit map.
3029
3030 3. Brought error texts together as macro definitions; clarified some of them;
3031 fixed one that was wrong - it said "range out of order" when it meant "invalid
3032 escape sequence".
3033
3034 4. Changed some char * arguments to const char *.
3035
3036 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3037
3038 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3039 pcretest.
3040
3041
3042 Version 0.96 16-Oct-97
3043 ----------------------
3044
3045 1. Added a simple "pgrep" utility to the distribution.
3046
3047 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3048 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3049 where "ddd" means "one or more decimal digits".
3050
3051 3. Fixed serious bug. If a pattern had a back reference, but the call to
3052 pcre_exec() didn't supply a large enough ovector to record the related
3053 identifying subpattern, the match always failed. PCRE now remembers the number
3054 of the largest back reference, and gets some temporary memory in which to save
3055 the offsets during matching if necessary, in order to ensure that
3056 backreferences always work.
3057
3058 4. Increased the compatibility with Perl in a number of ways:
3059
3060 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3061 to request this handling. The option can be set at compile or exec time.
3062
3063 (b) $ matches before a terminating newline by default; an option
3064 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3065 mode). The option can be set at compile or exec time.
3066
3067 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3068 the same as Perl's. If the decimal number it represents is less than 10
3069 or there aren't that many previous left capturing parentheses, an octal
3070 escape is read. Inside a character class, it's always an octal escape,
3071 even if it is a single digit.
3072
3073 (d) An escaped but undefined alphabetic character is taken as a literal,
3074 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3075 escapes.
3076
3077 (e) {0} is now permitted. (The previous item is removed from the compiled
3078 pattern).
3079
3080 5. Changed all the names of code files so that the basic parts are no longer
3081 than 10 characters, and abolished the teeny "globals.c" file.
3082
3083 6. Changed the handling of character classes; they are now done with a 32-byte
3084 bit map always.
3085
3086 7. Added the -d and /D options to pcretest to make it possible to look at the
3087 internals of compilation without having to recompile pcre.
3088
3089
3090 Version 0.95 23-Sep-97
3091 ----------------------
3092
3093 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3094 \x20 at the start of a run of normal characters. These were being treated as
3095 real characters, instead of the source characters being re-checked.
3096
3097
3098 Version 0.94 18-Sep-97
3099 ----------------------
3100
3101 1. The functions are now thread-safe, with the caveat that the global variables
3102 containing pointers to malloc() and free() or alternative functions are the
3103 same for all threads.
3104
3105 2. Get pcre_study() to generate a bitmap of initial characters for non-
3106 anchored patterns when this is possible, and use it if passed to pcre_exec().
3107
3108
3109 Version 0.93 15-Sep-97
3110 ----------------------
3111
3112 1. /(b)|(:+)/ was computing an incorrect first character.
3113
3114 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3115 but not actually doing anything yet.
3116
3117 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3118 as Perl does (e.g. [-az] or [az-]).
3119
3120 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3121 all possible positions.
3122
3123 5. Split up into different modules to avoid including unneeded functions in a
3124 compiled binary. However, compile and exec are still in one module. The "study"
3125 function is split off.
3126
3127 6. The character tables are now in a separate module whose source is generated
3128 by an auxiliary program - but can then be edited by hand if required. There are
3129 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3130 toupper() in the code.
3131
3132 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3133 make them global. Abolish the function for setting them, as the caller can now
3134 set them directly.
3135
3136
3137 Version 0.92 11-Sep-97
3138 ----------------------
3139
3140 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3141 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3142
3143 2. Caseless matching was not working in character classes if the characters in
3144 the pattern were in upper case.
3145
3146 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3147
3148 4. Make PCRE_ANCHORED public and accept as a compile option.
3149
3150 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3151 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3152 pass them.
3153
3154 6. Give an error if bad option bits passed at compile or run time.
3155
3156 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3157 pcretest to cause it to pass that flag.
3158
3159 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3160 options, and the first character, if set.
3161
3162 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3163
3164
3165 Version 0.91 10-Sep-97
3166 ----------------------
3167
3168 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3169 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3170
3171 2. PCRE was looping on encountering an indefinitely repeated back reference to
3172 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3173 Perl does - treats the match as successful.
3174
3175 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12