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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 389 - (show annotations) (download)
Sun Mar 15 18:24:05 2009 UTC (5 years, 8 months ago) by ph10
File size: 162441 byte(s)
Add PCRE_NO_START_OPTIMIZE

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.9 xx-xxx-09
5 ---------------------
6
7 1. When building with support for bzlib/zlib (pcregrep) and/or readline
8 (pcretest), all targets were linked against these libraries. This included
9 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
10 libraries. This caused unwanted dependencies to be created. This problem
11 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
12 pcretest is linked with readline.
13
14 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
15 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
16 moved outside it again, because FALSE and TRUE are already defined in AIX,
17 but BOOL is not.
18
19 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
20 PCRE_MATCH_LIMIT_RETURSION values as ints, when they should be long ints.
21
22 4. The pcregrep documentation said spaces were inserted as well as colons (or
23 hyphens) following file names and line numbers when outputting matching
24 lines. This is not true; no spaces are inserted. I have also clarified the
25 wording for the --colour (or --color) option.
26
27 5. In pcregrep, when --colour was used with -o, the list of matching strings
28 was not coloured; this is different to GNU grep, so I have changed it to be
29 the same.
30
31 6. When --colo(u)r was used in pcregrep, only the first matching substring in
32 each matching line was coloured. Now it goes on to look for further matches
33 of any of the test patterns, which is the same behaviour as GNU grep.
34
35 7. A pattern that could match an empty string could cause pcregrep to loop; it
36 doesn't make sense to accept an empty string match in pcregrep, so I have
37 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
38 seems to be how GNU grep behaves.
39
40 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
41 start or after a newline", because the conditional assertion was not being
42 correctly handled. The rule now is that both the assertion and what follows
43 in the first alternative must satisfy the test.
44
45 9. If auto-callout was enabled in a pattern with a conditional group, PCRE
46 could crash during matching.
47
48 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
49 used for matching.
50
51 11. Unicode property support in character classes was not working for
52 characters (bytes) greater than 127 when not in UTF-8 mode.
53
54 12. Added the -M command line option to pcretest.
55
56 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
57
58 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
59
60
61 Version 7.8 05-Sep-08
62 ---------------------
63
64 1. Replaced UCP searching code with optimized version as implemented for Ad
65 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
66 stage table and inline lookup instead of a function, giving speed ups of 2
67 to 5 times on some simple patterns that I tested. Permission was given to
68 distribute the MultiStage2.py script that generates the tables (it's not in
69 the tarball, but is in the Subversion repository).
70
71 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
72 scripts.
73
74 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
75 a group with a zero qualifier. The result of the study could be incorrect,
76 or the function might crash, depending on the pattern.
77
78 4. Caseless matching was not working for non-ASCII characters in back
79 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
80 It now works when Unicode Property Support is available.
81
82 5. In pcretest, an escape such as \x{de} in the data was always generating
83 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
84 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
85 truncation.
86
87 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
88
89 7. Added two (int) casts to pcregrep when printing the difference of two
90 pointers, in case they are 64-bit values.
91
92 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
93 test 2 if it fails.
94
95 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
96 and a #define of that name to empty if it is not externally set. This is to
97 allow users of MSVC to set it if necessary.
98
99 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
100 the convenience functions in the pcre_get.c source file.
101
102 11. An option change at the start of a pattern that had top-level alternatives
103 could cause overwriting and/or a crash. This command provoked a crash in
104 some environments:
105
106 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
107
108 This potential security problem was recorded as CVE-2008-2371.
109
110 12. For a pattern where the match had to start at the beginning or immediately
111 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
112 pcre_dfa_exec() could read past the end of the passed subject if there was
113 no match. To help with detecting such bugs (e.g. with valgrind), I modified
114 pcretest so that it places the subject at the end of its malloc-ed buffer.
115
116 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
117 exec() might read past the end of the data buffer in UTF-8 mode.
118
119 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
120 the data contained the byte 0x85 as part of a UTF-8 character within its
121 first line. This applied both to normal and DFA matching.
122
123 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
124 /^[^d]*?$/8 failed to match "abc".
125
126 16. Added a missing copyright notice to pcrecpp_internal.h.
127
128 17. Make it more clear in the documentation that values returned from
129 pcre_exec() in ovector are byte offsets, not character counts.
130
131 18. Tidied a few places to stop certain compilers from issuing warnings.
132
133 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
134 supplied by Stefan Weber. I made a further small update for 7.8 because
135 there is a change of source arrangements: the pcre_searchfuncs.c module is
136 replaced by pcre_ucd.c.
137
138
139 Version 7.7 07-May-08
140 ---------------------
141
142 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
143 a string to a long long, pretend we don't even have a long long." This is
144 done by checking for the strtoq, strtoll, and _strtoi64 functions.
145
146 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
147 pre-7.6 versions, which defined a global no_arg variable instead of putting
148 it in the RE class. (See also #8 below.)
149
150 3. Remove a line of dead code, identified by coverity and reported by Nuno
151 Lopes.
152
153 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
154
155 (1) The include/exclude patterns were being applied to the whole pathnames
156 of files, instead of just to the final components.
157
158 (2) If there was more than one level of directory, the subdirectories were
159 skipped unless they satisfied the include/exclude conditions. This is
160 inconsistent with GNU grep (and could even be seen as contrary to the
161 pcregrep specification - which I improved to make it absolutely clear).
162 The action now is always to scan all levels of directory, and just
163 apply the include/exclude patterns to regular files.
164
165 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
166 --exclude_dir in the tests to avoid scanning .svn directories.
167
168 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
169 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
170 doesn't support NULs in patterns.
171
172 7. Added some missing "const"s to declarations of static tables in
173 pcre_compile.c and pcre_dfa_exec.c.
174
175 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
176 caused by fix #2 above. (Subsequently also a second patch to fix the
177 first patch. And a third patch - this was a messy problem.)
178
179 9. Applied Craig's patch to remove the use of push_back().
180
181 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
182 matching function regexec().
183
184 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
185 which, however, unlike Perl's \g{...}, are subroutine calls, not back
186 references. PCRE supports relative numbers with this syntax (I don't think
187 Oniguruma does).
188
189 12. Previously, a group with a zero repeat such as (...){0} was completely
190 omitted from the compiled regex. However, this means that if the group
191 was called as a subroutine from elsewhere in the pattern, things went wrong
192 (an internal error was given). Such groups are now left in the compiled
193 pattern, with a new opcode that causes them to be skipped at execution
194 time.
195
196 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
197 to the way PCRE behaves:
198
199 (a) A lone ] character is dis-allowed (Perl treats it as data).
200
201 (b) A back reference to an unmatched subpattern matches an empty string
202 (Perl fails the current match path).
203
204 (c) A data ] in a character class must be notated as \] because if the
205 first data character in a class is ], it defines an empty class. (In
206 Perl it is not possible to have an empty class.) The empty class []
207 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
208 The negative empty class [^] matches any one character, independently
209 of the DOTALL setting.
210
211 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
212 non-existent subpattern following a character class starting with ']' and
213 containing () gave an internal compiling error instead of "reference to
214 non-existent subpattern". Fortunately, when the pattern did exist, the
215 compiled code was correct. (When scanning forwards to check for the
216 existencd of the subpattern, it was treating the data ']' as terminating
217 the class, so got the count wrong. When actually compiling, the reference
218 was subsequently set up correctly.)
219
220 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
221 it was being rejected as not supported by pcre_dfa_exec(), even though
222 other assertions are supported. I have made pcre_dfa_exec() support
223 (*FAIL).
224
225 16. The implementation of 13c above involved the invention of a new opcode,
226 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
227 cannot be changed at match time, I realized I could make a small
228 improvement to matching performance by compiling OP_ALLANY instead of
229 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
230 on the OP_ANY path.
231
232 17. Compiling pcretest on Windows with readline support failed without the
233 following two fixes: (1) Make the unistd.h include conditional on
234 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
235
236 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
237 ncurses library to be included for pcretest when ReadLine support is
238 requested, but also to allow for it to be overridden. This patch came from
239 Daniel Bergström.
240
241 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
242 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
243 any errors with the current Unicode tables. Thanks to Peter Kankowski for
244 spotting this.
245
246
247 Version 7.6 28-Jan-08
248 ---------------------
249
250 1. A character class containing a very large number of characters with
251 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
252 overflow.
253
254 2. Patch to cut out the "long long" test in pcrecpp_unittest when
255 HAVE_LONG_LONG is not defined.
256
257 3. Applied Christian Ehrlicher's patch to update the CMake build files to
258 bring them up to date and include new features. This patch includes:
259
260 - Fixed PH's badly added libz and libbz2 support.
261 - Fixed a problem with static linking.
262 - Added pcredemo. [But later removed - see 7 below.]
263 - Fixed dftables problem and added an option.
264 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
265 HAVE_LONG_LONG.
266 - Added readline support for pcretest.
267 - Added an listing of the option settings after cmake has run.
268
269 4. A user submitted a patch to Makefile that makes it easy to create
270 "pcre.dll" under mingw when using Configure/Make. I added stuff to
271 Makefile.am that cause it to include this special target, without
272 affecting anything else. Note that the same mingw target plus all
273 the other distribution libraries and programs are now supported
274 when configuring with CMake (see 6 below) instead of with
275 Configure/Make.
276
277 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
278 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
279 exported in the Windows port". It has not yet been confirmed that the patch
280 solves the problem, but it does no harm.
281
282 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
283 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
284 with CMake, and also correct the comment about stack recursion.
285
286 7. Remove the automatic building of pcredemo from the ./configure system and
287 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
288 of a program that users should build themselves after PCRE is installed, so
289 building it automatically is not really right. What is more, it gave
290 trouble in some build environments.
291
292 8. Further tidies to CMakeLists.txt from Sheri and Christian.
293
294
295 Version 7.5 10-Jan-08
296 ---------------------
297
298 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
299 values in parens when parsing an RE using the C++ wrapper."
300
301 2. Negative specials like \S did not work in character classes in UTF-8 mode.
302 Characters greater than 255 were excluded from the class instead of being
303 included.
304
305 3. The same bug as (2) above applied to negated POSIX classes such as
306 [:^space:].
307
308 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
309 defined or documented. It seems to have been a typo for PCRE_STATIC, so
310 I have changed it.
311
312 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
313 first named subpattern) and a construct such as (?&a) would reference the
314 first named subpattern whose name started with "a" (in other words, the
315 length check was missing). Both these problems are fixed. "Subpattern name
316 expected" is now given for (?&) (a zero-length name), and this patch also
317 makes it give the same error for \k'' (previously it complained that that
318 was a reference to a non-existent subpattern).
319
320 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
321 this is right because (?- can be followed by option settings as well as by
322 digits. I have, however, made the messages clearer.
323
324 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
325 than the number used in the conditional) now cause a compile-time error.
326 This is actually not compatible with Perl, which accepts such patterns, but
327 treats the conditional as always being FALSE (as PCRE used to), but it
328 seems to me that giving a diagnostic is better.
329
330 8. Change "alphameric" to the more common word "alphanumeric" in comments
331 and messages.
332
333 9. Fix two occurrences of "backslash" in comments that should have been
334 "backspace".
335
336 10. Remove two redundant lines of code that can never be obeyed (their function
337 was moved elsewhere).
338
339 11. The program that makes PCRE's Unicode character property table had a bug
340 which caused it to generate incorrect table entries for sequences of
341 characters that have the same character type, but are in different scripts.
342 It amalgamated them into a single range, with the script of the first of
343 them. In other words, some characters were in the wrong script. There were
344 thirteen such cases, affecting characters in the following ranges:
345
346 U+002b0 - U+002c1
347 U+0060c - U+0060d
348 U+0061e - U+00612
349 U+0064b - U+0065e
350 U+0074d - U+0076d
351 U+01800 - U+01805
352 U+01d00 - U+01d77
353 U+01d9b - U+01dbf
354 U+0200b - U+0200f
355 U+030fc - U+030fe
356 U+03260 - U+0327f
357 U+0fb46 - U+0fbb1
358 U+10450 - U+1049d
359
360 12. The -o option (show only the matching part of a line) for pcregrep was not
361 compatible with GNU grep in that, if there was more than one match in a
362 line, it showed only the first of them. It now behaves in the same way as
363 GNU grep.
364
365 13. If the -o and -v options were combined for pcregrep, it printed a blank
366 line for every non-matching line. GNU grep prints nothing, and pcregrep now
367 does the same. The return code can be used to tell if there were any
368 non-matching lines.
369
370 14. Added --file-offsets and --line-offsets to pcregrep.
371
372 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
373 infinitely looping recursion. The bug was that positive lookaheads were not
374 being skipped when checking for a possible empty match (negative lookaheads
375 and both kinds of lookbehind were skipped).
376
377 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
378 inclusion of <windows.h> to before rather than after the definition of
379 INVALID_FILE_ATTRIBUTES (patch from David Byron).
380
381 17. Specifying a possessive quantifier with a specific limit for a Unicode
382 character property caused pcre_compile() to compile bad code, which led at
383 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
384 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
385 caused the error; without that there was no problem.
386
387 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
388
389 19. Added --enable-pcretest-libreadline.
390
391 20. In pcrecpp.cc, the variable 'count' was incremented twice in
392 RE::GlobalReplace(). As a result, the number of replacements returned was
393 double what it should be. I removed one of the increments, but Craig sent a
394 later patch that removed the other one (the right fix) and added unit tests
395 that check the return values (which was not done before).
396
397 21. Several CMake things:
398
399 (1) Arranged that, when cmake is used on Unix, the libraries end up with
400 the names libpcre and libpcreposix, not just pcre and pcreposix.
401
402 (2) The above change means that pcretest and pcregrep are now correctly
403 linked with the newly-built libraries, not previously installed ones.
404
405 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
406
407 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
408 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
409 UTF-8 newline character). The key issue is that the pattern starts .*;
410 this means that the match must be either at the beginning, or after a
411 newline. The bug was in the code for advancing after a failed match and
412 checking that the new position followed a newline. It was not taking
413 account of UTF-8 characters correctly.
414
415 23. PCRE was behaving differently from Perl in the way it recognized POSIX
416 character classes. PCRE was not treating the sequence [:...:] as a
417 character class unless the ... were all letters. Perl, however, seems to
418 allow any characters between [: and :], though of course it rejects as
419 unknown any "names" that contain non-letters, because all the known class
420 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
421 for example, whereas PCRE did not - it did not recognize a POSIX character
422 class. This seemed a bit dangerous, so the code has been changed to be
423 closer to Perl. The behaviour is not identical to Perl, because PCRE will
424 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
425 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
426 Perl does, and where it didn't before.
427
428 24. Rewrite so as to remove the single use of %n from pcregrep because in some
429 Windows environments %n is disabled by default.
430
431
432 Version 7.4 21-Sep-07
433 ---------------------
434
435 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
436 means that a class such as [\s] counted as "explicit reference to CR or
437 LF". That isn't really right - the whole point of the change was to try to
438 help when there was an actual mention of one of the two characters. So now
439 the change happens only if \r or \n (or a literal CR or LF) character is
440 encountered.
441
442 2. The 32-bit options word was also used for 6 internal flags, but the numbers
443 of both had grown to the point where there were only 3 bits left.
444 Fortunately, there was spare space in the data structure, and so I have
445 moved the internal flags into a new 16-bit field to free up more option
446 bits.
447
448 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
449 but did not set the internal JCHANGED flag - either of these is enough to
450 control the way the "get" function works - but the PCRE_INFO_JCHANGED
451 facility is supposed to tell if (?J) was ever used, so now (?J) at the
452 start sets both bits.
453
454 4. Added options (at build time, compile time, exec time) to change \R from
455 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
456
457 5. doc/pcresyntax.html was missing from the distribution.
458
459 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
460 compatibility, even though it is no longer used.
461
462 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
463 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
464 windows.h file is present (where different names are used). [This was
465 reversed later after testing - see 16 below.]
466
467 8. Changed all #include <config.h> to #include "config.h". There were also
468 some further <pcre.h> cases that I changed to "pcre.h".
469
470 9. When pcregrep was used with the --colour option, it missed the line ending
471 sequence off the lines that it output.
472
473 10. It was pointed out to me that arrays of string pointers cause lots of
474 relocations when a shared library is dynamically loaded. A technique of
475 using a single long string with a table of offsets can drastically reduce
476 these. I have refactored PCRE in four places to do this. The result is
477 dramatic:
478
479 Originally: 290
480 After changing UCP table: 187
481 After changing error message table: 43
482 After changing table of "verbs" 36
483 After changing table of Posix names 22
484
485 Thanks to the folks working on Gregex for glib for this insight.
486
487 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
488 unicode-properties was also set.
489
490 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
491
492 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
493 checked only for CRLF.
494
495 14. Added casts to pcretest.c to avoid compiler warnings.
496
497 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
498
499 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
500 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
501 entirely. This removes changes made in 7 above.
502
503 17. The CMake files have been updated, and there is now more information about
504 building with CMake in the NON-UNIX-USE document.
505
506
507 Version 7.3 28-Aug-07
508 ---------------------
509
510 1. In the rejigging of the build system that eventually resulted in 7.1, the
511 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
512 brackets there is not right, since it causes compilers to look for an
513 installed pcre.h, not the version that is in the source that is being
514 compiled (which of course may be different). I have changed it back to:
515
516 #include "pcre.h"
517
518 I have a vague recollection that the change was concerned with compiling in
519 different directories, but in the new build system, that is taken care of
520 by the VPATH setting the Makefile.
521
522 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
523 when the subject happened to end in the byte 0x85 (e.g. if the last
524 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
525 characters but of course it shouldn't be taken as a newline when it is part
526 of another character. The bug was that, for an unlimited repeat of . in
527 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
528 characters when looking for a newline.
529
530 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
531
532 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
533 in debug output.
534
535 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
536 long printing in the pcrecpp unittest when running under MinGW.
537
538 6. ESC_K was left out of the EBCDIC table.
539
540 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
541 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
542 limit also applies to "virtual nesting" when a pattern is recursive, and in
543 this case 1000 isn't so big. I have been able to remove this limit at the
544 expense of backing off one optimization in certain circumstances. Normally,
545 when pcre_exec() would call its internal match() function recursively and
546 immediately return the result unconditionally, it uses a "tail recursion"
547 feature to save stack. However, when a subpattern that can match an empty
548 string has an unlimited repetition quantifier, it no longer makes this
549 optimization. That gives it a stack frame in which to save the data for
550 checking that an empty string has been matched. Previously this was taken
551 from the 1000-entry workspace that had been reserved. So now there is no
552 explicit limit, but more stack is used.
553
554 8. Applied Daniel's patches to solve problems with the import/export magic
555 syntax that is required for Windows, and which was going wrong for the
556 pcreposix and pcrecpp parts of the library. These were overlooked when this
557 problem was solved for the main library.
558
559 9. There were some crude static tests to avoid integer overflow when computing
560 the size of patterns that contain repeated groups with explicit upper
561 limits. As the maximum quantifier is 65535, the maximum group length was
562 set at 30,000 so that the product of these two numbers did not overflow a
563 32-bit integer. However, it turns out that people want to use groups that
564 are longer than 30,000 bytes (though not repeat them that many times).
565 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
566 made it possible to implement the integer overflow checks in a much more
567 dynamic way, which I have now done. The artificial limitation on group
568 length has been removed - we now have only the limit on the total length of
569 the compiled pattern, which depends on the LINK_SIZE setting.
570
571 10. Fixed a bug in the documentation for get/copy named substring when
572 duplicate names are permitted. If none of the named substrings are set, the
573 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
574 empty string.
575
576 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
577 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
578 because the ] is interpreted as the first data character and the
579 terminating ] is not found. PCRE has been made compatible with Perl in this
580 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
581 cause memory overwriting.
582
583 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
584 string has been matched (to stop an infinite loop). It was not recognizing
585 a conditional subpattern that could match an empty string if that
586 subpattern was within another subpattern. For example, it looped when
587 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
588 condition was not nested. This bug has been fixed.
589
590 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
591 past the start of the subject in the presence of bytes with the top bit
592 set, for example "\x8aBCD".
593
594 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
595 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
596
597 14. Optimized (?!) to (*FAIL).
598
599 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
600 This restricts code points to be within the range 0 to 0x10FFFF, excluding
601 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
602 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
603 does: it's just the validity check that is more restrictive.
604
605 16. Inserted checks for integer overflows during escape sequence (backslash)
606 processing, and also fixed erroneous offset values for syntax errors during
607 backslash processing.
608
609 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
610 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
611
612 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
613 caused an overrun.
614
615 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
616 something other than just ASCII characters) inside a group that had an
617 unlimited repeat caused a loop at compile time (while checking to see
618 whether the group could match an empty string).
619
620 20. Debugging a pattern containing \p or \P could cause a crash. For example,
621 [\P{Any}] did so. (Error in the code for printing property names.)
622
623 21. An orphan \E inside a character class could cause a crash.
624
625 22. A repeated capturing bracket such as (A)? could cause a wild memory
626 reference during compilation.
627
628 23. There are several functions in pcre_compile() that scan along a compiled
629 expression for various reasons (e.g. to see if it's fixed length for look
630 behind). There were bugs in these functions when a repeated \p or \P was
631 present in the pattern. These operators have additional parameters compared
632 with \d, etc, and these were not being taken into account when moving along
633 the compiled data. Specifically:
634
635 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
636 length.
637
638 (b) An item such as \pL+ within a repeated group could cause crashes or
639 loops.
640
641 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
642 "reference to non-existent subpattern" error.
643
644 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
645
646 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
647 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
648
649 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
650
651 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
652 character were causing crashes (broken optimization).
653
654 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
655 \p or \P) caused a compile-time loop.
656
657 28. More problems have arisen in unanchored patterns when CRLF is a valid line
658 break. For example, the unstudied pattern [\r\n]A does not match the string
659 "\r\nA" because change 7.0/46 below moves the current point on by two
660 characters after failing to match at the start. However, the pattern \nA
661 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
662 the same is true. There doesn't seem any very clean way out of this, but
663 what I have chosen to do makes the common cases work: PCRE now takes note
664 of whether there can be an explicit match for \r or \n anywhere in the
665 pattern, and if so, 7.0/46 no longer applies. As part of this change,
666 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
667 pattern has explicit CR or LF references.
668
669 29. Added (*CR) etc for changing newline setting at start of pattern.
670
671
672 Version 7.2 19-Jun-07
673 ---------------------
674
675 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
676 which is apparently normally available under Windows.
677
678 2. Re-jig the pcregrep tests with different newline settings in an attempt
679 to make them independent of the local environment's newline setting.
680
681 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
682
683 4. Some of the "internals" tests were previously cut out when the link size
684 was not 2, because the output contained actual offsets. The recent new
685 "Z" feature of pcretest means that these can be cut out, making the tests
686 usable with all link sizes.
687
688 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
689 stack recursion. This gives a massive performance boost under BSD, but just
690 a small improvement under Linux. However, it saves one field in the frame
691 in all cases.
692
693 6. Added more features from the forthcoming Perl 5.10:
694
695 (a) (?-n) (where n is a string of digits) is a relative subroutine or
696 recursion call. It refers to the nth most recently opened parentheses.
697
698 (b) (?+n) is also a relative subroutine call; it refers to the nth next
699 to be opened parentheses.
700
701 (c) Conditions that refer to capturing parentheses can be specified
702 relatively, for example, (?(-2)... or (?(+3)...
703
704 (d) \K resets the start of the current match so that everything before
705 is not part of it.
706
707 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
708
709 (f) \g{name} is another synonym - part of Perl 5.10's unification of
710 reference syntax.
711
712 (g) (?| introduces a group in which the numbering of parentheses in each
713 alternative starts with the same number.
714
715 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
716
717 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
718 PCRE_INFO_JCHANGED.
719
720 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
721 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
722 for detecting groups that can match an empty string.
723
724 9. A pattern with a very large number of alternatives (more than several
725 hundred) was running out of internal workspace during the pre-compile
726 phase, where pcre_compile() figures out how much memory will be needed. A
727 bit of new cunning has reduced the workspace needed for groups with
728 alternatives. The 1000-alternative test pattern now uses 12 bytes of
729 workspace instead of running out of the 4096 that are available.
730
731 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
732
733 11. Applied patch from Google to remove an optimization that didn't quite work.
734 The report of the bug said:
735
736 pcrecpp::RE("a*").FullMatch("aaa") matches, while
737 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
738 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
739
740 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
741 it matched the wrong number of bytes.
742
743
744 Version 7.1 24-Apr-07
745 ---------------------
746
747 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
748 that is more "standard", making use of automake and other Autotools. There
749 is some re-arrangement of the files and adjustment of comments consequent
750 on this.
751
752 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
753 for recursive directory scanning broke on some systems because the files
754 are not scanned in any specific order and on different systems the order
755 was different. A call to "sort" has been inserted into RunGrepTest for the
756 approprate test as a short-term fix. In the longer term there may be an
757 alternative.
758
759 3. I had an email from Eric Raymond about problems translating some of PCRE's
760 man pages to HTML (despite the fact that I distribute HTML pages, some
761 people do their own conversions for various reasons). The problems
762 concerned the use of low-level troff macros .br and .in. I have therefore
763 removed all such uses from the man pages (some were redundant, some could
764 be replaced by .nf/.fi pairs). The 132html script that I use to generate
765 HTML has been updated to handle .nf/.fi and to complain if it encounters
766 .br or .in.
767
768 4. Updated comments in configure.ac that get placed in config.h.in and also
769 arranged for config.h to be included in the distribution, with the name
770 config.h.generic, for the benefit of those who have to compile without
771 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
772
773 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
774 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
775 makevp.bat for the current PCRE, using the additional files
776 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
777
778 6. A Windows user reported a minor discrepancy with test 2, which turned out
779 to be caused by a trailing space on an input line that had got lost in his
780 copy. The trailing space was an accident, so I've just removed it.
781
782 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
783 that is needed.
784
785 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
786 as "const" (a) because they are and (b) because it helps the PHP
787 maintainers who have recently made a script to detect big data structures
788 in the php code that should be moved to the .rodata section. I remembered
789 to update Builducptable as well, so it won't revert if ucptable.h is ever
790 re-created.
791
792 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
793 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
794 order to be able to cut out the UTF-8 tables in the latter when UTF-8
795 support is not required. This saves 1.5-2K of code, which is important in
796 some applications.
797
798 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
799 so as not to refer to the tables, even though these functions will never be
800 called when UTF-8 support is disabled. Otherwise there are problems with a
801 shared library.
802
803 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
804
805 (a) It was defining its arguments as char * instead of void *.
806
807 (b) It was assuming that all moves were upwards in memory; this was true
808 a long time ago when I wrote it, but is no longer the case.
809
810 The emulated memove() is provided for those environments that have neither
811 memmove() nor bcopy(). I didn't think anyone used it these days, but that
812 is clearly not the case, as these two bugs were recently reported.
813
814 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
815 and Detrail to create the HTML documentation, the .txt form of the man
816 pages, and it removes trailing spaces from listed files. It also creates
817 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
818 case, it wraps all the #defines with #ifndefs. This script should be run
819 before "make dist".
820
821 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
822 with Unicode property support.
823
824 (a) For a maximizing quantifier, if the two different cases of the
825 character were of different lengths in their UTF-8 codings (there are
826 some cases like this - I found 11), and the matching function had to
827 back up over a mixture of the two cases, it incorrectly assumed they
828 were both the same length.
829
830 (b) When PCRE was configured to use the heap rather than the stack for
831 recursion during matching, it was not correctly preserving the data for
832 the other case of a UTF-8 character when checking ahead for a match
833 while processing a minimizing repeat. If the check also involved
834 matching a wide character, but failed, corruption could cause an
835 erroneous result when trying to check for a repeat of the original
836 character.
837
838 13. Some tidying changes to the testing mechanism:
839
840 (a) The RunTest script now detects the internal link size and whether there
841 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
842 values substituted by "configure". (The RunGrepTest script already did
843 this for UTF-8.) The configure.ac script no longer substitutes the
844 relevant variables.
845
846 (b) The debugging options /B and /D in pcretest show the compiled bytecode
847 with length and offset values. This means that the output is different
848 for different internal link sizes. Test 2 is skipped for link sizes
849 other than 2 because of this, bypassing the problem. Unfortunately,
850 there was also a test in test 3 (the locale tests) that used /B and
851 failed for link sizes other than 2. Rather than cut the whole test out,
852 I have added a new /Z option to pcretest that replaces the length and
853 offset values with spaces. This is now used to make test 3 independent
854 of link size. (Test 2 will be tidied up later.)
855
856 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
857 segmentation fault instead of returning the appropriate error message.
858
859 15. In multiline mode when the newline sequence was set to "any", the pattern
860 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
861 This doesn't seem right; it now treats the CRLF combination as the line
862 ending, and so does not match in that case. It's only a pattern such as ^$
863 that would hit this one: something like ^ABC$ would have failed after \r
864 and then tried again after \r\n.
865
866 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
867 in an attempt to make files that differ only in their line terminators
868 compare equal. This works on Linux.
869
870 17. Under certain error circumstances pcregrep might try to free random memory
871 as it exited. This is now fixed, thanks to valgrind.
872
873 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
874 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
875 was because its rules for how to advance for /g after matching an empty
876 string at the end of a line did not allow for this case. They now check for
877 it specially.
878
879 20. pcretest is supposed to handle patterns and data of any length, by
880 extending its buffers when necessary. It was getting this wrong when the
881 buffer for a data line had to be extended.
882
883 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
884 CRLF as a newline sequence.
885
886 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
887 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
888 I have nevertheless tidied it up.
889
890 23. Added some casts to kill warnings from HP-UX ia64 compiler.
891
892 24. Added a man page for pcre-config.
893
894
895 Version 7.0 19-Dec-06
896 ---------------------
897
898 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
899 moving to gcc 4.1.1.
900
901 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
902 sys/time.h, which is documented as needed for this function. It doesn't
903 seem to matter on Linux, but it showed up on some releases of OS X.
904
905 3. It seems that there are systems where bytes whose values are greater than
906 127 match isprint() in the "C" locale. The "C" locale should be the
907 default when a C program starts up. In most systems, only ASCII printing
908 characters match isprint(). This difference caused the output from pcretest
909 to vary, making some of the tests fail. I have changed pcretest so that:
910
911 (a) When it is outputting text in the compiled version of a pattern, bytes
912 other than 32-126 are always shown as hex escapes.
913
914 (b) When it is outputting text that is a matched part of a subject string,
915 it does the same, unless a different locale has been set for the match
916 (using the /L modifier). In this case, it uses isprint() to decide.
917
918 4. Fixed a major bug that caused incorrect computation of the amount of memory
919 required for a compiled pattern when options that changed within the
920 pattern affected the logic of the preliminary scan that determines the
921 length. The relevant options are -x, and -i in UTF-8 mode. The result was
922 that the computed length was too small. The symptoms of this bug were
923 either the PCRE error "internal error: code overflow" from pcre_compile(),
924 or a glibc crash with a message such as "pcretest: free(): invalid next
925 size (fast)". Examples of patterns that provoked this bug (shown in
926 pcretest format) are:
927
928 /(?-x: )/x
929 /(?x)(?-x: \s*#\s*)/
930 /((?i)[\x{c0}])/8
931 /(?i:[\x{c0}])/8
932
933 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
934 is now done differently.
935
936 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
937 wrapper classes; (b) implement a new function in the C++ scanner that is
938 more efficient than the old way of doing things because it avoids levels of
939 recursion in the regex matching; (c) add a paragraph to the documentation
940 for the FullMatch() function.
941
942 6. The escape sequence \n was being treated as whatever was defined as
943 "newline". Not only was this contrary to the documentation, which states
944 that \n is character 10 (hex 0A), but it also went horribly wrong when
945 "newline" was defined as CRLF. This has been fixed.
946
947 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
948 was being set to -1 for the "end of line" case (supposedly a value that no
949 character can have). Though this value is never used (the check for end of
950 line is "zero bytes in current character"), it caused compiler complaints.
951 I've changed it to 0xffffffff.
952
953 8. In pcre_version.c, the version string was being built by a sequence of
954 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
955 string (as it is for production releases) called a macro with an empty
956 argument. The C standard says the result of this is undefined. The gcc
957 compiler treats it as an empty string (which was what was wanted) but it is
958 reported that Visual C gives an error. The source has been hacked around to
959 avoid this problem.
960
961 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
962 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
963 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
964 of them did).
965
966 10. Originally, pcretest opened its input and output without "b"; then I was
967 told that "b" was needed in some environments, so it was added for release
968 5.0 to both the input and output. (It makes no difference on Unix-like
969 systems.) Later I was told that it is wrong for the input on Windows. I've
970 now abstracted the modes into two macros, to make it easier to fiddle with
971 them, and removed "b" from the input mode under Windows.
972
973 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
974
975 12. Added -help and --help to pcretest as an official way of being reminded
976 of the options.
977
978 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
979 and pcrecpp.cc because they annoy compilers at high warning levels.
980
981 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
982
983 15. Fixed an occurrence of == in configure.ac that should have been = (shell
984 scripts are not C programs :-) and which was not noticed because it works
985 on Linux.
986
987 16. pcretest is supposed to handle any length of pattern and data line (as one
988 line or as a continued sequence of lines) by extending its input buffer if
989 necessary. This feature was broken for very long pattern lines, leading to
990 a string of junk being passed to pcre_compile() if the pattern was longer
991 than about 50K.
992
993 17. I have done a major re-factoring of the way pcre_compile() computes the
994 amount of memory needed for a compiled pattern. Previously, there was code
995 that made a preliminary scan of the pattern in order to do this. That was
996 OK when PCRE was new, but as the facilities have expanded, it has become
997 harder and harder to keep it in step with the real compile phase, and there
998 have been a number of bugs (see for example, 4 above). I have now found a
999 cunning way of running the real compile function in a "fake" mode that
1000 enables it to compute how much memory it would need, while actually only
1001 ever using a few hundred bytes of working memory and without too many
1002 tests of the mode. This should make future maintenance and development
1003 easier. A side effect of this work is that the limit of 200 on the nesting
1004 depth of parentheses has been removed (though this was never a serious
1005 limitation, I suspect). However, there is a downside: pcre_compile() now
1006 runs more slowly than before (30% or more, depending on the pattern). I
1007 hope this isn't a big issue. There is no effect on runtime performance.
1008
1009 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1010 newline (only possible for the last line of a file) and it was a
1011 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1012
1013 19. Added additional timing features to pcretest. (1) The -tm option now times
1014 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1015 separate command line item, by a number that specifies the number of
1016 repeats to use when timing. The default is 50000; this gives better
1017 precision, but takes uncomfortably long for very large patterns.
1018
1019 20. Extended pcre_study() to be more clever in cases where a branch of a
1020 subpattern has no definite first character. For example, (a*|b*)[cd] would
1021 previously give no result from pcre_study(). Now it recognizes that the
1022 first character must be a, b, c, or d.
1023
1024 21. There was an incorrect error "recursive call could loop indefinitely" if
1025 a subpattern (or the entire pattern) that was being tested for matching an
1026 empty string contained only one non-empty item after a nested subpattern.
1027 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1028 incorrectly, because the \d was being skipped in the check.
1029
1030 22. The pcretest program now has a new pattern option /B and a command line
1031 option -b, which is equivalent to adding /B to every pattern. This causes
1032 it to show the compiled bytecode, without the additional information that
1033 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1034 is the same as /B/I).
1035
1036 23. A new optimization is now able automatically to treat some sequences such
1037 as a*b as a*+b. More specifically, if something simple (such as a character
1038 or a simple class like \d) has an unlimited quantifier, and is followed by
1039 something that cannot possibly match the quantified thing, the quantifier
1040 is automatically "possessified".
1041
1042 24. A recursive reference to a subpattern whose number was greater than 39
1043 went wrong under certain circumstances in UTF-8 mode. This bug could also
1044 have affected the operation of pcre_study().
1045
1046 25. Realized that a little bit of performance could be had by replacing
1047 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1048
1049 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1050
1051 27. Possessive quantifiers such as a++ were previously implemented by turning
1052 them into atomic groups such as ($>a+). Now they have their own opcodes,
1053 which improves performance. This includes the automatically created ones
1054 from 23 above.
1055
1056 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1057 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1058 the first matched character to be a colon. This applied both to named and
1059 numbered groups.
1060
1061 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1062
1063 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1064 building PCRE on Windows easier, so I have included it in the distribution.
1065
1066 31. There is now a check in pcretest against a ridiculously large number being
1067 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1068 loop, the loop is abandoned.
1069
1070 32. Forward references to subpatterns in conditions such as (?(2)...) where
1071 subpattern 2 is defined later cause pcre_compile() to search forwards in
1072 the pattern for the relevant set of parentheses. This search went wrong
1073 when there were unescaped parentheses in a character class, parentheses
1074 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1075
1076 33. "Subroutine" calls and backreferences were previously restricted to
1077 referencing subpatterns earlier in the regex. This restriction has now
1078 been removed.
1079
1080 34. Added a number of extra features that are going to be in Perl 5.10. On the
1081 whole, these are just syntactic alternatives for features that PCRE had
1082 previously implemented using the Python syntax or my own invention. The
1083 other formats are all retained for compatibility.
1084
1085 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1086 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1087 also .NET compatible.
1088
1089 (b) A recursion or subroutine call to a named group can now be defined as
1090 (?&name) as well as (?P>name).
1091
1092 (c) A backreference to a named group can now be defined as \k<name> or
1093 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1094 5.10, are also .NET compatible.
1095
1096 (d) A conditional reference to a named group can now use the syntax
1097 (?(<name>) or (?('name') as well as (?(name).
1098
1099 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1100 groups (named and numbered) that are never evaluated inline, but can be
1101 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1102 is always false. There may be only one alternative in such a group.
1103
1104 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1105 as the simple (?(R). The condition is true only if the most recent
1106 recursion is that of the given number or name. It does not search out
1107 through the entire recursion stack.
1108
1109 (g) The escape \gN or \g{N} has been added, where N is a positive or
1110 negative number, specifying an absolute or relative reference.
1111
1112 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1113 some "unreachable code" warnings.
1114
1115 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1116 things, this adds five new scripts.
1117
1118 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1119 There were also incompatibilities regarding the handling of \Q..\E inside
1120 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1121 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1122
1123 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1124 matches an empty string, and forcibly breaks the loop. There were bugs in
1125 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1126 against aaaa the result was just "a" rather than "aaaa", for example. Two
1127 separate and independent bugs (that affected different cases) have been
1128 fixed.
1129
1130 39. Refactored the code to abolish the use of different opcodes for small
1131 capturing bracket numbers. This is a tidy that I avoided doing when I
1132 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1133 The new approach is not only tidier, it makes it possible to reduce the
1134 memory needed to fix the previous bug (38).
1135
1136 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1137 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1138 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1139 mode.
1140
1141 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1142 report.
1143
1144 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1145 copy construction and assignment in the C++ wrapper.
1146
1147 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1148 couple of bugs that could have given wrong results in the "--newline=crlf"
1149 case.
1150
1151 44. Added a number of casts and did some reorganization of signed/unsigned int
1152 variables following suggestions from Dair Grant. Also renamed the variable
1153 "this" as "item" because it is a C++ keyword.
1154
1155 45. Arranged for dftables to add
1156
1157 #include "pcre_internal.h"
1158
1159 to pcre_chartables.c because without it, gcc 4.x may remove the array
1160 definition from the final binary if PCRE is built into a static library and
1161 dead code stripping is activated.
1162
1163 46. For an unanchored pattern, if a match attempt fails at the start of a
1164 newline sequence, and the newline setting is CRLF or ANY, and the next two
1165 characters are CRLF, advance by two characters instead of one.
1166
1167
1168 Version 6.7 04-Jul-06
1169 ---------------------
1170
1171 1. In order to handle tests when input lines are enormously long, pcretest has
1172 been re-factored so that it automatically extends its buffers when
1173 necessary. The code is crude, but this _is_ just a test program. The
1174 default size has been increased from 32K to 50K.
1175
1176 2. The code in pcre_study() was using the value of the re argument before
1177 testing it for NULL. (Of course, in any sensible call of the function, it
1178 won't be NULL.)
1179
1180 3. The memmove() emulation function in pcre_internal.h, which is used on
1181 systems that lack both memmove() and bcopy() - that is, hardly ever -
1182 was missing a "static" storage class specifier.
1183
1184 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1185 containing an extended class (one that cannot be represented by a bitmap
1186 because it contains high-valued characters or Unicode property items, e.g.
1187 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1188 pattern, but PCRE should not loop if you do not (it no longer does).
1189 [Detail: two cases were found: (a) a repeated subpattern containing an
1190 extended class; (b) a recursive reference to a subpattern that followed a
1191 previous extended class. It wasn't skipping over the extended class
1192 correctly when UTF-8 mode was not set.]
1193
1194 5. A negated single-character class was not being recognized as fixed-length
1195 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1196 compile error "lookbehind assertion is not fixed length".
1197
1198 6. The RunPerlTest auxiliary script was showing an unexpected difference
1199 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1200 write a Perl script that can interpret lines of an input file either as
1201 byte characters or as UTF-8, which is what "perltest" was being required to
1202 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1203 can't do is switch easily at run time between having the "use utf8;" pragma
1204 or not. In the end, I fudged it by using the RunPerlTest script to insert
1205 "use utf8;" explicitly for the UTF-8 tests.
1206
1207 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1208 the end of the subject string, contrary to the documentation and to what
1209 Perl does. This was true of both matching functions. Now it matches only at
1210 the start of the subject and immediately after *internal* newlines.
1211
1212 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1213 a pointer to an int instead of a pointer to an unsigned long int. This
1214 caused problems on 64-bit systems.
1215
1216 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1217 instance of the 'standard' template library not being so standard".
1218
1219 10. There was no check on the number of named subpatterns nor the maximum
1220 length of a subpattern name. The product of these values is used to compute
1221 the size of the memory block for a compiled pattern. By supplying a very
1222 long subpattern name and a large number of named subpatterns, the size
1223 computation could be caused to overflow. This is now prevented by limiting
1224 the length of names to 32 characters, and the number of named subpatterns
1225 to 10,000.
1226
1227 11. Subpatterns that are repeated with specific counts have to be replicated in
1228 the compiled pattern. The size of memory for this was computed from the
1229 length of the subpattern and the repeat count. The latter is limited to
1230 65535, but there was no limit on the former, meaning that integer overflow
1231 could in principle occur. The compiled length of a repeated subpattern is
1232 now limited to 30,000 bytes in order to prevent this.
1233
1234 12. Added the optional facility to have named substrings with the same name.
1235
1236 13. Added the ability to use a named substring as a condition, using the
1237 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1238 are numbers (not recommended). Forward references are permitted.
1239
1240 14. Added forward references in named backreferences (if you see what I mean).
1241
1242 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1243 pattern could run off the end of the subject. For example, the pattern
1244 "(?s)(.{1,5})"8 did this with the subject "ab".
1245
1246 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1247 PCRE_CASELESS was set when matching characters that were quantified with ?
1248 or *.
1249
1250 17. A character class other than a single negated character that had a minimum
1251 but no maximum quantifier - for example [ab]{6,} - was not handled
1252 correctly by pce_dfa_exec(). It would match only one character.
1253
1254 18. A valid (though odd) pattern that looked like a POSIX character
1255 class but used an invalid character after [ (for example [[,abc,]]) caused
1256 pcre_compile() to give the error "Failed: internal error: code overflow" or
1257 in some cases to crash with a glibc free() error. This could even happen if
1258 the pattern terminated after [[ but there just happened to be a sequence of
1259 letters, a binary zero, and a closing ] in the memory that followed.
1260
1261 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1262 over the years. Originally (before any Unicode support), just the bottom 8
1263 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1264 output from "man perlunicode" includes this:
1265
1266 The regular expression compiler produces polymorphic opcodes. That
1267 is, the pattern adapts to the data and automatically switches to
1268 the Unicode character scheme when presented with Unicode data--or
1269 instead uses a traditional byte scheme when presented with byte
1270 data.
1271
1272 Sadly, a wide octal escape does not cause a switch, and in a string with
1273 no other multibyte characters, these octal escapes are treated as before.
1274 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1275 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1276 Unicode string.
1277
1278 I have not perpetrated such confusion in PCRE. Up till now, it took just
1279 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1280 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1281 translate to the appropriate multibyte character.
1282
1283 29. Applied some refactoring to reduce the number of warnings from Microsoft
1284 and Borland compilers. This has included removing the fudge introduced
1285 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1286 a warning about an unused variable.
1287
1288 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1289 characters since release 4.0, because Perl (from release 5.004) does not.
1290 [Or at least, is documented not to: some releases seem to be in conflict
1291 with the documentation.] However, when a pattern was studied with
1292 pcre_study() and all its branches started with \s, PCRE still included VT
1293 as a possible starting character. Of course, this did no harm; it just
1294 caused an unnecessary match attempt.
1295
1296 22. Removed a now-redundant internal flag bit that recorded the fact that case
1297 dependency changed within the pattern. This was once needed for "required
1298 byte" processing, but is no longer used. This recovers a now-scarce options
1299 bit. Also moved the least significant internal flag bit to the most-
1300 significant bit of the word, which was not previously used (hangover from
1301 the days when it was an int rather than a uint) to free up another bit for
1302 the future.
1303
1304 23. Added support for CRLF line endings as well as CR and LF. As well as the
1305 default being selectable at build time, it can now be changed at runtime
1306 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1307 specify that it is scanning data with non-default line endings.
1308
1309 24. Changed the definition of CXXLINK to make it agree with the definition of
1310 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1311
1312 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1313 recursions. This makes a big different to stack usage for some patterns.
1314
1315 26. If a subpattern containing a named recursion or subroutine reference such
1316 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1317 the space required for the compiled pattern went wrong and gave too small a
1318 value. Depending on the environment, this could lead to "Failed: internal
1319 error: code overflow at offset 49" or "glibc detected double free or
1320 corruption" errors.
1321
1322 27. Applied patches from Google (a) to support the new newline modes and (b) to
1323 advance over multibyte UTF-8 characters in GlobalReplace.
1324
1325 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1326 difference for some implementation of PCRE in some Windows version.
1327
1328 29. Added some extra testing facilities to pcretest:
1329
1330 \q<number> in a data line sets the "match limit" value
1331 \Q<number> in a data line sets the "match recursion limt" value
1332 -S <number> sets the stack size, where <number> is in megabytes
1333
1334 The -S option isn't available for Windows.
1335
1336
1337 Version 6.6 06-Feb-06
1338 ---------------------
1339
1340 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1341 in pcreposix.h. I have copied the definition from pcre.h.
1342
1343 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1344 because pcre.h is no longer a built file.
1345
1346 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1347 not normally included in the compiled code.
1348
1349
1350 Version 6.5 01-Feb-06
1351 ---------------------
1352
1353 1. When using the partial match feature with pcre_dfa_exec(), it was not
1354 anchoring the second and subsequent partial matches at the new starting
1355 point. This could lead to incorrect results. For example, with the pattern
1356 /1234/, partially matching against "123" and then "a4" gave a match.
1357
1358 2. Changes to pcregrep:
1359
1360 (a) All non-match returns from pcre_exec() were being treated as failures
1361 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1362 error message is output. Some extra information is given for the
1363 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1364 probably the only errors that are likely to be caused by users (by
1365 specifying a regex that has nested indefinite repeats, for instance).
1366 If there are more than 20 of these errors, pcregrep is abandoned.
1367
1368 (b) A binary zero was treated as data while matching, but terminated the
1369 output line if it was written out. This has been fixed: binary zeroes
1370 are now no different to any other data bytes.
1371
1372 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1373 used to set a locale for matching. The --locale=xxxx long option has
1374 been added (no short equivalent) to specify a locale explicitly on the
1375 pcregrep command, overriding the environment variables.
1376
1377 (d) When -B was used with -n, some line numbers in the output were one less
1378 than they should have been.
1379
1380 (e) Added the -o (--only-matching) option.
1381
1382 (f) If -A or -C was used with -c (count only), some lines of context were
1383 accidentally printed for the final match.
1384
1385 (g) Added the -H (--with-filename) option.
1386
1387 (h) The combination of options -rh failed to suppress file names for files
1388 that were found from directory arguments.
1389
1390 (i) Added the -D (--devices) and -d (--directories) options.
1391
1392 (j) Added the -F (--fixed-strings) option.
1393
1394 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1395
1396 (l) Added the --colo(u)r option.
1397
1398 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1399 is not present by default.
1400
1401 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1402 items such as (?R) or (?1), when the recursion could match a number of
1403 alternatives. If it matched one of the alternatives, but subsequently,
1404 outside the recursion, there was a failure, the code tried to back up into
1405 the recursion. However, because of the way PCRE is implemented, this is not
1406 possible, and the result was an incorrect result from the match.
1407
1408 In order to prevent this happening, the specification of recursion has
1409 been changed so that all such subpatterns are automatically treated as
1410 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1411
1412 4. I had overlooked the fact that, in some locales, there are characters for
1413 which isalpha() is true but neither isupper() nor islower() are true. In
1414 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1415 and ordfeminine) are like this. This affected the treatment of \w and \W
1416 when they appeared in character classes, but not when they appeared outside
1417 a character class. The bit map for "word" characters is now created
1418 separately from the results of isalnum() instead of just taking it from the
1419 upper, lower, and digit maps. (Plus the underscore character, of course.)
1420
1421 5. The above bug also affected the handling of POSIX character classes such as
1422 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1423 permanent tables. Instead, the bit maps for such a class were previously
1424 created as the appropriate unions of the upper, lower, and digit bitmaps.
1425 Now they are created by subtraction from the [[:word:]] class, which has
1426 its own bitmap.
1427
1428 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1429 It is created by subtracting the vertical space characters (\x09, \x0a,
1430 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1431 subtraction was done in the overall bitmap for a character class, meaning
1432 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1433 be recognized. This bug has been fixed.
1434
1435 7. Patches from the folks at Google:
1436
1437 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1438 real life, but is still worth protecting against".
1439
1440 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1441 regular expressions".
1442
1443 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1444 have it.
1445
1446 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1447 "configure" and the latter not, in order to fix a problem somebody had
1448 with compiling the Arg class on HP-UX.
1449
1450 (e) Improve the error-handling of the C++ wrapper a little bit.
1451
1452 (f) New tests for checking recursion limiting.
1453
1454 8. The pcre_memmove() function, which is used only if the environment does not
1455 have a standard memmove() function (and is therefore rarely compiled),
1456 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1457 returning a result (though PCRE never actually uses the result).
1458
1459 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1460 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1461 returned instead of calling malloc() with an overflowing number that would
1462 most likely cause subsequent chaos.
1463
1464 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1465
1466 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1467 with this option is matched, the nmatch and pmatch options of regexec() are
1468 ignored.
1469
1470 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1471 provided in case anyone wants to the the POSIX interface with UTF-8
1472 strings.
1473
1474 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1475 C++ linking (needed for some HP-UX environments).
1476
1477 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1478 (unused parameter) and in the pcre_printint() function (omitted "default"
1479 switch label when the default is to do nothing).
1480
1481 15. Added some code to make it possible, when PCRE is compiled as a C++
1482 library, to replace subject pointers for pcre_exec() with a smart pointer
1483 class, thus making it possible to process discontinuous strings.
1484
1485 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1486 much the same function. They were added by different people who were trying
1487 to make PCRE easy to compile on non-Unix systems. It has been suggested
1488 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1489 for compiling on Windows systems. I have therefore replaced it with
1490 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1491 defaults to "extern" for C or "extern C" for C++, which works fine on
1492 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1493 SCOPE with something explicit in config.h. In addition:
1494
1495 (a) pcreposix.h still had just "extern" instead of either of these macros;
1496 I have replaced it with PCRE_DATA_SCOPE.
1497
1498 (b) Functions such as _pcre_xclass(), which are internal to the library,
1499 but external in the C sense, all had PCRE_EXPORT in their definitions.
1500 This is apparently wrong for the Windows case, so I have removed it.
1501 (It makes no difference on Unix-like systems.)
1502
1503 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1504 of recursive calls to match(). This is different to MATCH_LIMIT because
1505 that limits the total number of calls to match(), not all of which increase
1506 the depth of recursion. Limiting the recursion depth limits the amount of
1507 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1508 when PCRE is compiled, and changed at run time. A patch from Google adds
1509 this functionality to the C++ interface.
1510
1511 18. Changes to the handling of Unicode character properties:
1512
1513 (a) Updated the table to Unicode 4.1.0.
1514
1515 (b) Recognize characters that are not in the table as "Cn" (undefined).
1516
1517 (c) I revised the way the table is implemented to a much improved format
1518 which includes recognition of ranges. It now supports the ranges that
1519 are defined in UnicodeData.txt, and it also amalgamates other
1520 characters into ranges. This has reduced the number of entries in the
1521 table from around 16,000 to around 3,000, thus reducing its size
1522 considerably. I realized I did not need to use a tree structure after
1523 all - a binary chop search is just as efficient. Having reduced the
1524 number of entries, I extended their size from 6 bytes to 8 bytes to
1525 allow for more data.
1526
1527 (d) Added support for Unicode script names via properties such as \p{Han}.
1528
1529 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1530 matching that character.
1531
1532 20. When matching a repeated Unicode property with a minimum greater than zero,
1533 (for example \pL{2,}), PCRE could look past the end of the subject if it
1534 reached it while seeking the minimum number of characters. This could
1535 happen only if some of the characters were more than one byte long, because
1536 there is a check for at least the minimum number of bytes.
1537
1538 21. Refactored the implementation of \p and \P so as to be more general, to
1539 allow for more different types of property in future. This has changed the
1540 compiled form incompatibly. Anybody with saved compiled patterns that use
1541 \p or \P will have to recompile them.
1542
1543 22. Added "Any" and "L&" to the supported property types.
1544
1545 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1546 but give a compile time error if the value is greater than 0xff.
1547
1548 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1549 accidentally not being installed or uninstalled.
1550
1551 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1552 made were to insert the current release number. This seemed silly, because
1553 it made things harder for people building PCRE on systems that don't run
1554 "configure". I have turned pcre.h into a distributed file, no longer built
1555 by "configure", with the version identification directly included. There is
1556 no longer a pcre.h.in file.
1557
1558 However, this change necessitated a change to the pcre-config script as
1559 well. It is built from pcre-config.in, and one of the substitutions was the
1560 release number. I have updated configure.ac so that ./configure now finds
1561 the release number by grepping pcre.h.
1562
1563 26. Added the ability to run the tests under valgrind.
1564
1565
1566 Version 6.4 05-Sep-05
1567 ---------------------
1568
1569 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1570 "--" to be printed when multiple files were scanned, even when none of the
1571 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1572 consider it to be a bug, and have restored the previous behaviour.
1573
1574 2. A couple of code tidies to get rid of compiler warnings.
1575
1576 3. The pcretest program used to cheat by referring to symbols in the library
1577 whose names begin with _pcre_. These are internal symbols that are not
1578 really supposed to be visible externally, and in some environments it is
1579 possible to suppress them. The cheating is now confined to including
1580 certain files from the library's source, which is a bit cleaner.
1581
1582 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1583 file's purpose clearer.
1584
1585 5. Reorganized pcre_ucp_findchar().
1586
1587
1588 Version 6.3 15-Aug-05
1589 ---------------------
1590
1591 1. The file libpcre.pc.in did not have general read permission in the tarball.
1592
1593 2. There were some problems when building without C++ support:
1594
1595 (a) If C++ support was not built, "make install" and "make test" still
1596 tried to test it.
1597
1598 (b) There were problems when the value of CXX was explicitly set. Some
1599 changes have been made to try to fix these, and ...
1600
1601 (c) --disable-cpp can now be used to explicitly disable C++ support.
1602
1603 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1604 backslash in a target when C++ was disabled. This confuses some
1605 versions of "make", apparently. Using an intermediate variable solves
1606 this. (Same for CPP_LOBJ.)
1607
1608 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1609 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1610 necessary on certain architectures.
1611
1612 4. Added a setting of -export-symbols-regex to the link command to remove
1613 those symbols that are exported in the C sense, but actually are local
1614 within the library, and not documented. Their names all begin with
1615 "_pcre_". This is not a perfect job, because (a) we have to except some
1616 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1617 available (and never for static libraries). I have made a note to try to
1618 find a way round (a) in the future.
1619
1620
1621 Version 6.2 01-Aug-05
1622 ---------------------
1623
1624 1. There was no test for integer overflow of quantifier values. A construction
1625 such as {1111111111111111} would give undefined results. What is worse, if
1626 a minimum quantifier for a parenthesized subpattern overflowed and became
1627 negative, the calculation of the memory size went wrong. This could have
1628 led to memory overwriting.
1629
1630 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1631
1632 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1633 operating environments where this matters.
1634
1635 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1636 PCRE options from within the C++ wrapper.
1637
1638 5. Named capturing subpatterns were not being correctly counted when a pattern
1639 was compiled. This caused two problems: (a) If there were more than 100
1640 such subpatterns, the calculation of the memory needed for the whole
1641 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1642 back references of the form \12, where the number was greater than 9, were
1643 not recognized as back references, even though there were sufficient
1644 previous subpatterns.
1645
1646 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1647 versions of gcc, e.g. 2.95.4.
1648
1649
1650 Version 6.1 21-Jun-05
1651 ---------------------
1652
1653 1. There was one reference to the variable "posix" in pcretest.c that was not
1654 surrounded by "#if !defined NOPOSIX".
1655
1656 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1657 the cross-check on the old pcre_info() function, for the benefit of the
1658 cut-down version of PCRE that is currently imported into Exim.
1659
1660 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1661 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1662 patterns that start (?i) but I don't think that matters. The use of (?i) is
1663 just an example; this all applies to the other options as well.
1664
1665 4. Since libtool seems to echo the compile commands it is issuing, the output
1666 from "make" can be reduced a bit by putting "@" in front of each libtool
1667 compile command.
1668
1669 5. Patch from the folks at Google for configure.in to be a bit more thorough
1670 in checking for a suitable C++ installation before trying to compile the
1671 C++ stuff. This should fix a reported problem when a compiler was present,
1672 but no suitable headers.
1673
1674 6. The man pages all had just "PCRE" as their title. I have changed them to
1675 be the relevant file name. I have also arranged that these names are
1676 retained in the file doc/pcre.txt, which is a concatenation in text format
1677 of all the man pages except the little individual ones for each function.
1678
1679 7. The NON-UNIX-USE file had not been updated for the different set of source
1680 files that come with release 6. I also added a few comments about the C++
1681 wrapper.
1682
1683
1684 Version 6.0 07-Jun-05
1685 ---------------------
1686
1687 1. Some minor internal re-organization to help with my DFA experiments.
1688
1689 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1690 didn't matter for the library itself when fully configured, but did matter
1691 when compiling without UCP support, or within Exim, where the ucp files are
1692 not imported.
1693
1694 3. Refactoring of the library code to split up the various functions into
1695 different source modules. The addition of the new DFA matching code (see
1696 below) to a single monolithic source would have made it really too
1697 unwieldy, quite apart from causing all the code to be include in a
1698 statically linked application, when only some functions are used. This is
1699 relevant even without the DFA addition now that patterns can be compiled in
1700 one application and matched in another.
1701
1702 The downside of splitting up is that there have to be some external
1703 functions and data tables that are used internally in different modules of
1704 the library but which are not part of the API. These have all had their
1705 names changed to start with "_pcre_" so that they are unlikely to clash
1706 with other external names.
1707
1708 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1709 a different (DFA) algorithm. Although it is slower than the original
1710 function, it does have some advantages for certain types of matching
1711 problem.
1712
1713 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1714 including restarting after a partial match.
1715
1716 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1717 defined when compiling for Windows was sent to me. I have put it into the
1718 code, though I have no means of testing or verifying it.
1719
1720 7. Added the pcre_refcount() auxiliary function.
1721
1722 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1723 match before or at the first newline in the subject string. In pcretest,
1724 the /f option on a pattern can be used to set this.
1725
1726 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1727 would behave wrongly. This has been present in PCRE since release 4.0.
1728
1729 10. A number of changes to the pcregrep command:
1730
1731 (a) Refactored how -x works; insert ^(...)$ instead of setting
1732 PCRE_ANCHORED and checking the length, in preparation for adding
1733 something similar for -w.
1734
1735 (b) Added the -w (match as a word) option.
1736
1737 (c) Refactored the way lines are read and buffered so as to have more
1738 than one at a time available.
1739
1740 (d) Implemented a pcregrep test script.
1741
1742 (e) Added the -M (multiline match) option. This allows patterns to match
1743 over several lines of the subject. The buffering ensures that at least
1744 8K, or the rest of the document (whichever is the shorter) is available
1745 for matching (and similarly the previous 8K for lookbehind assertions).
1746
1747 (f) Changed the --help output so that it now says
1748
1749 -w, --word-regex(p)
1750
1751 instead of two lines, one with "regex" and the other with "regexp"
1752 because that confused at least one person since the short forms are the
1753 same. (This required a bit of code, as the output is generated
1754 automatically from a table. It wasn't just a text change.)
1755
1756 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1757 option but starts with a hyphen. Could be a pattern or a path name
1758 starting with a hyphen, for instance.
1759
1760 (h) "-" can be given as a file name to represent stdin.
1761
1762 (i) When file names are being printed, "(standard input)" is used for
1763 the standard input, for compatibility with GNU grep. Previously
1764 "<stdin>" was used.
1765
1766 (j) The option --label=xxx can be used to supply a name to be used for
1767 stdin when file names are being printed. There is no short form.
1768
1769 (k) Re-factored the options decoding logic because we are going to add
1770 two more options that take data. Such options can now be given in four
1771 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1772
1773 (l) Added the -A, -B, and -C options for requesting that lines of context
1774 around matches be printed.
1775
1776 (m) Added the -L option to print the names of files that do not contain
1777 any matching lines, that is, the complement of -l.
1778
1779 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1780 continue to scan other files.
1781
1782 (o) The -s option was incorrectly implemented. For compatibility with other
1783 greps, it now suppresses the error message for a non-existent or non-
1784 accessible file (but not the return code). There is a new option called
1785 -q that suppresses the output of matching lines, which was what -s was
1786 previously doing.
1787
1788 (p) Added --include and --exclude options to specify files for inclusion
1789 and exclusion when recursing.
1790
1791 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1792 Hopefully, it now does.
1793
1794 12. Missing cast in pcre_study().
1795
1796 13. Added an "uninstall" target to the makefile.
1797
1798 14. Replaced "extern" in the function prototypes in Makefile.in with
1799 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1800 world, but is set differently for Windows.
1801
1802 15. Added a second compiling function called pcre_compile2(). The only
1803 difference is that it has an extra argument, which is a pointer to an
1804 integer error code. When there is a compile-time failure, this is set
1805 non-zero, in addition to the error test pointer being set to point to an
1806 error message. The new argument may be NULL if no error number is required
1807 (but then you may as well call pcre_compile(), which is now just a
1808 wrapper). This facility is provided because some applications need a
1809 numeric error indication, but it has also enabled me to tidy up the way
1810 compile-time errors are handled in the POSIX wrapper.
1811
1812 16. Added VPATH=.libs to the makefile; this should help when building with one
1813 prefix path and installing with another. (Or so I'm told by someone who
1814 knows more about this stuff than I do.)
1815
1816 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1817 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1818 match everything, including newlines. This is not POSIX-compatible, but
1819 somebody wanted the feature. From pcretest it can be activated by using
1820 both the P and the s flags.
1821
1822 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1823
1824 19. libpcre.pc was being incorrectly installed as executable.
1825
1826 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1827 it now also looks for '\r' so that it will work unmodified on Windows.
1828
1829 21. Added Google's contributed C++ wrapper to the distribution.
1830
1831 22. Added some untidy missing memory free() calls in pcretest, to keep
1832 Electric Fence happy when testing.
1833
1834
1835
1836 Version 5.0 13-Sep-04
1837 ---------------------
1838
1839 1. Internal change: literal characters are no longer packed up into items
1840 containing multiple characters in a single byte-string. Each character
1841 is now matched using a separate opcode. However, there may be more than one
1842 byte in the character in UTF-8 mode.
1843
1844 2. The pcre_callout_block structure has two new fields: pattern_position and
1845 next_item_length. These contain the offset in the pattern to the next match
1846 item, and its length, respectively.
1847
1848 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1849 insertion of callouts before each pattern item. Added the /C option to
1850 pcretest to make use of this.
1851
1852 4. On the advice of a Windows user, the lines
1853
1854 #if defined(_WIN32) || defined(WIN32)
1855 _setmode( _fileno( stdout ), 0x8000 );
1856 #endif /* defined(_WIN32) || defined(WIN32) */
1857
1858 have been added to the source of pcretest. This apparently does useful
1859 magic in relation to line terminators.
1860
1861 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1862 for the benefit of those environments where the "b" makes a difference.
1863
1864 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1865 to know about it. I have put a hack into configure.in that adds in code
1866 to set GCC=yes if CC=icc. This seems to end up at a point in the
1867 generated configure script that is early enough to affect the setting of
1868 compiler options, which is what is needed, but I have no means of testing
1869 whether it really works. (The user who reported this had patched the
1870 generated configure script, which of course I cannot do.)
1871
1872 LATER: After change 22 below (new libtool files), the configure script
1873 seems to know about icc (and also ecc). Therefore, I have commented out
1874 this hack in configure.in.
1875
1876 7. Added support for pkg-config (2 patches were sent in).
1877
1878 8. Negated POSIX character classes that used a combination of internal tables
1879 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1880 [[:^ascii]]. Typically, they would match almost any characters. The other
1881 POSIX classes were not broken in this way.
1882
1883 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1884 to find the match, as PCRE was deluded into thinking that the match had to
1885 start at the start point or following a newline. The same bug applied to
1886 patterns with negative forward assertions or any backward assertions
1887 preceding ".*" at the start, unless the pattern required a fixed first
1888 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1889
1890 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1891 starting at the last subject character, bytes beyond the end of the subject
1892 string were read.
1893
1894 11. Renamed the variable "class" as "classbits" to make life easier for C++
1895 users. (Previously there was a macro definition, but it apparently wasn't
1896 enough.)
1897
1898 12. Added the new field "tables" to the extra data so that tables can be passed
1899 in at exec time, or the internal tables can be re-selected. This allows
1900 a compiled regex to be saved and re-used at a later time by a different
1901 program that might have everything at different addresses.
1902
1903 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1904 -R library as well as a -L library.
1905
1906 14. The debugging options of pcretest (-d on the command line or D on a
1907 pattern) showed incorrect output for anything following an extended class
1908 that contained multibyte characters and which was followed by a quantifier.
1909
1910 15. Added optional support for general category Unicode character properties
1911 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1912 support. It adds about 90K to the size of the library. The meanings of the
1913 inbuilt class escapes such as \d and \s have NOT been changed.
1914
1915 16. Updated pcredemo.c to include calls to free() to release the memory for the
1916 compiled pattern.
1917
1918 17. The generated file chartables.c was being created in the source directory
1919 instead of in the building directory. This caused the build to fail if the
1920 source directory was different from the building directory, and was
1921 read-only.
1922
1923 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1924 file. No doubt somebody will tell me if they don't make sense... Also added
1925 Dan Mooney's comments about building on OpenVMS.
1926
1927 19. Added support for partial matching via the PCRE_PARTIAL option for
1928 pcre_exec() and the \P data escape in pcretest.
1929
1930 20. Extended pcretest with 3 new pattern features:
1931
1932 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1933 write the compiled pattern to the file whose name is "rest-of-line".
1934 This is a straight binary dump of the data, with the saved pointer to
1935 the character tables forced to be NULL. The study data, if any, is
1936 written too. After writing, pcretest reads a new pattern.
1937
1938 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1939 compiled pattern from the given file. There must not be any
1940 occurrences of "<" in the file name (pretty unlikely); if there are,
1941 pcretest will instead treat the initial "<" as a pattern delimiter.
1942 After reading in the pattern, pcretest goes on to read data lines as
1943 usual.
1944
1945 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1946 and 16-bit fields in a compiled pattern, to simulate a pattern that
1947 was compiled on a host of opposite endianness.
1948
1949 21. The pcre-exec() function can now cope with patterns that were compiled on
1950 hosts of opposite endianness, with this restriction:
1951
1952 As for any compiled expression that is saved and used later, the tables
1953 pointer field cannot be preserved; the extra_data field in the arguments
1954 to pcre_exec() should be used to pass in a tables address if a value
1955 other than the default internal tables were used at compile time.
1956
1957 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1958 now diagnosed as an error. Previously, most of the time, a negative number
1959 would have been treated as zero, but if in addition "ovector" was passed as
1960 NULL, a crash could occur.
1961
1962 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1963 new versions from the libtool 1.5 distribution (the last one is a copy of
1964 a file called libtool.m4). This seems to have fixed the need to patch
1965 "configure" to support Darwin 1.3 (which I used to do). However, I still
1966 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1967 workstation).
1968
1969 24. Changed the PCRE licence to be the more standard "BSD" licence.
1970
1971
1972 Version 4.5 01-Dec-03
1973 ---------------------
1974
1975 1. There has been some re-arrangement of the code for the match() function so
1976 that it can be compiled in a version that does not call itself recursively.
1977 Instead, it keeps those local variables that need separate instances for
1978 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1979 needs to "recurse". Keeping track of where control must go is done by means
1980 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1981 hide most of the details from the main code, and operates only if
1982 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1983 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1984 operating.
1985
1986 To make it easier for callers to provide specially tailored get/free
1987 functions for this usage, two new functions, pcre_stack_malloc, and
1988 pcre_stack_free, are used. They are always called in strict stacking order,
1989 and the size of block requested is always the same.
1990
1991 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1992 PCRE has been compiled to use the stack or the heap for recursion. The
1993 -C option of pcretest uses this to show which version is compiled.
1994
1995 A new data escape \S, is added to pcretest; it causes the amounts of store
1996 obtained and freed by both kinds of malloc/free at match time to be added
1997 to the output.
1998
1999 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2000 what's available on my current Linux desktop machine.
2001
2002 3. When matching a UTF-8 string, the test for a valid string at the start has
2003 been extended. If start_offset is not zero, PCRE now checks that it points
2004 to a byte that is the start of a UTF-8 character. If not, it returns
2005 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2006 this is necessary because there may be backward assertions in the pattern.
2007 When matching the same subject several times, it may save resources to use
2008 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2009
2010 4. The code for checking the validity of UTF-8 strings has been tightened so
2011 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2012 containing "overlong sequences".
2013
2014 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2015 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2016 should have been "&", but it just so happened that all the cases this let
2017 through by mistake were picked up later in the function.
2018
2019 6. I had used a variable called "isblank" - this is a C99 function, causing
2020 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2021
2022 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2023 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2024 so the version is shown; (c) stop "make test" echoing "./RunTest".
2025
2026 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2027
2028 9. The prototype for memmove() for systems that don't have it was using
2029 size_t, but the inclusion of the header that defines size_t was later. I've
2030 moved the #includes for the C headers earlier to avoid this.
2031
2032 10. Added some adjustments to the code to make it easier to compiler on certain
2033 special systems:
2034
2035 (a) Some "const" qualifiers were missing.
2036 (b) Added the macro EXPORT before all exported functions; by default this
2037 is defined to be empty.
2038 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2039 that it reads its output file name as an argument instead of writing
2040 to the standard output and assuming this can be redirected.
2041
2042 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2043 class containing characters with values greater than 255, PCRE compilation
2044 went into a loop.
2045
2046 12. A recursive reference to a subpattern that was within another subpattern
2047 that had a minimum quantifier of zero caused PCRE to crash. For example,
2048 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2049 recursion. If the recursively-called subpattern itself had a zero repeat,
2050 that was OK.
2051
2052 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2053 buffer into which it was copied (for escape processing) was still set at
2054 1024, so long lines caused crashes.
2055
2056 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2057 "internal error: code overflow...". This applied to any character class
2058 that was followed by a possessive quantifier.
2059
2060 15. Modified the Makefile to add libpcre.la as a prerequisite for
2061 libpcreposix.la because I was told this is needed for a parallel build to
2062 work.
2063
2064 16. If a pattern that contained .* following optional items at the start was
2065 studied, the wrong optimizing data was generated, leading to matching
2066 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2067 matching string must start with a or b or c. The correct conclusion for
2068 this pattern is that a match can start with any character.
2069
2070
2071 Version 4.4 13-Aug-03
2072 ---------------------
2073
2074 1. In UTF-8 mode, a character class containing characters with values between
2075 127 and 255 was not handled correctly if the compiled pattern was studied.
2076 In fixing this, I have also improved the studying algorithm for such
2077 classes (slightly).
2078
2079 2. Three internal functions had redundant arguments passed to them. Removal
2080 might give a very teeny performance improvement.
2081
2082 3. Documentation bug: the value of the capture_top field in a callout is *one
2083 more than* the number of the hightest numbered captured substring.
2084
2085 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2086 in incorrectly linking with a previously installed version. They now link
2087 explicitly with libpcre.la.
2088
2089 5. configure.in no longer needs to recognize Cygwin specially.
2090
2091 6. A problem in pcre.in for Windows platforms is fixed.
2092
2093 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2094 pcretest, it used to include the size of the study block as part of its
2095 output. Unfortunately, the structure contains a field that has a different
2096 size on different hardware architectures. This meant that the tests that
2097 showed this size failed. As the block is currently always of a fixed size,
2098 this information isn't actually particularly useful in pcretest output, so
2099 I have just removed it.
2100
2101 8. Three pre-processor statements accidentally did not start in column 1.
2102 Sadly, there are *still* compilers around that complain, even though
2103 standard C has not required this for well over a decade. Sigh.
2104
2105 9. In pcretest, the code for checking callouts passed small integers in the
2106 callout_data field, which is a void * field. However, some picky compilers
2107 complained about the casts involved for this on 64-bit systems. Now
2108 pcretest passes the address of the small integer instead, which should get
2109 rid of the warnings.
2110
2111 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2112 both compile and run time, and gives an error if an invalid UTF-8 sequence
2113 is found. There is a option for disabling this check in cases where the
2114 string is known to be correct and/or the maximum performance is wanted.
2115
2116 11. In response to a bug report, I changed one line in Makefile.in from
2117
2118 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2119 to
2120 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2121
2122 to look similar to other lines, but I have no way of telling whether this
2123 is the right thing to do, as I do not use Windows. No doubt I'll get told
2124 if it's wrong...
2125
2126
2127 Version 4.3 21-May-03
2128 ---------------------
2129
2130 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2131 Makefile.
2132
2133 2. Some refactoring to improve the quality of the code:
2134
2135 (i) The utf8_table... variables are now declared "const".
2136
2137 (ii) The code for \cx, which used the "case flipping" table to upper case
2138 lower case letters, now just substracts 32. This is ASCII-specific,
2139 but the whole concept of \cx is ASCII-specific, so it seems
2140 reasonable.
2141
2142 (iii) PCRE was using its character types table to recognize decimal and
2143 hexadecimal digits in the pattern. This is silly, because it handles
2144 only 0-9, a-f, and A-F, but the character types table is locale-
2145 specific, which means strange things might happen. A private
2146 table is now used for this - though it costs 256 bytes, a table is
2147 much faster than multiple explicit tests. Of course, the standard
2148 character types table is still used for matching digits in subject
2149 strings against \d.
2150
2151 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2152 ending in _t are). So I've renamed it as ESC_tee.
2153
2154 3. The first argument for regexec() in the POSIX wrapper should have been
2155 defined as "const".
2156
2157 4. Changed pcretest to use malloc() for its buffers so that they can be
2158 Electric Fenced for debugging.
2159
2160 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2161 to read one or more bytes before the start of the subject string. Often this
2162 had no effect on PCRE's behaviour, but in some circumstances it could
2163 provoke a segmentation fault.
2164
2165 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2166 to try to read one or more bytes before the start of the subject string.
2167
2168 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2169 UTF-8 support could misbehave in various ways if the subject string
2170 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2171 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2172 back over UTF-8 characters.)
2173
2174
2175 Version 4.2 14-Apr-03
2176 ---------------------
2177
2178 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2179
2180 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2181 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2182 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2183 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2184 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2185 and BUILD_EXEEXT
2186 Note: automatic setting of the BUILD variables is not yet working
2187 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2188 compile-time but not at link-time
2189 [LINK]: use for linking executables only
2190 make different versions for Windows and non-Windows
2191 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2192 libraries
2193 [LINK_FOR_BUILD]: new variable
2194 [OBJEXT]: use throughout
2195 [EXEEXT]: use throughout
2196 <winshared>: new target
2197 <wininstall>: new target
2198 <dftables.o>: use native compiler
2199 <dftables>: use native linker
2200 <install>: handle Windows platform correctly
2201 <clean>: ditto
2202 <check>: ditto
2203 copy DLL to top builddir before testing
2204
2205 As part of these changes, -no-undefined was removed again. This was reported
2206 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2207 in any case.
2208
2209 3. Some tidies to get rid of compiler warnings:
2210
2211 . In the match_data structure, match_limit was an unsigned long int, whereas
2212 match_call_count was an int. I've made them both unsigned long ints.
2213
2214 . In pcretest the fact that a const uschar * doesn't automatically cast to
2215 a void * provoked a warning.
2216
2217 . Turning on some more compiler warnings threw up some "shadow" variables
2218 and a few more missing casts.
2219
2220 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2221 option, a class that contained a single character with a value between 128
2222 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2223
2224 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2225 option, a class that contained several characters, but with at least one
2226 whose value was between 128 and 255 caused PCRE to crash.
2227
2228
2229 Version 4.1 12-Mar-03
2230 ---------------------
2231
2232 1. Compiling with gcc -pedantic found a couple of places where casts were
2233 needed, and a string in dftables.c that was longer than standard compilers are
2234 required to support.
2235
2236 2. Compiling with Sun's compiler found a few more places where the code could
2237 be tidied up in order to avoid warnings.
2238
2239 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2240 first of these names is deprecated in the latest Autoconf in favour of the name
2241 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2242 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2243 analogy I have changed it to CFLAGS_FOR_BUILD.
2244
2245 4. Added -no-undefined to the linking command in the Makefile, because this is
2246 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2247 linking step for the pcreposix library.
2248
2249 5. PCRE was failing to diagnose the case of two named groups with the same
2250 name.
2251
2252 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2253 literal character that is needed in the subject for a match, and scans along to
2254 ensure that it is present before embarking on the full matching process. This
2255 saves time in cases of nested unlimited repeats that are never going to match.
2256 Problem: the scan can take a lot of time if the subject is very long (e.g.
2257 megabytes), thus penalizing straightforward matches. It is now done only if the
2258 amount of subject to be scanned is less than 1000 bytes.
2259
2260 7. A lesser problem with the same optimization is that it was recording the
2261 first character of an anchored pattern as "needed", thus provoking a search
2262 right along the subject, even when the first match of the pattern was going to
2263 fail. The "needed" character is now not set for anchored patterns, unless it
2264 follows something in the pattern that is of non-fixed length. Thus, it still
2265 fulfils its original purpose of finding quick non-matches in cases of nested
2266 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2267
2268
2269 Version 4.0 17-Feb-03
2270 ---------------------
2271
2272 1. If a comment in an extended regex that started immediately after a meta-item
2273 extended to the end of string, PCRE compiled incorrect data. This could lead to
2274 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2275
2276 2. Moved to autoconf 2.53 and libtool 1.4.2.
2277
2278 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2279 the special perltest8 script is no longer needed - all the tests can be run
2280 from a single perltest script.
2281
2282 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2283 by \s. It has now been removed in PCRE. This means it isn't recognized as
2284 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2285 class [:space:] *does* include VT, thereby creating a mess.
2286
2287 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2288 space and tab.
2289
2290 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2291 its new features into the main test script, reducing the number of scripts.
2292
2293 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2294 were backward compatible, and made the (?i) apply to the whole pattern, as if
2295 /i were given. Now it behaves more logically, and applies the option setting
2296 only to what follows. PCRE has been changed to follow suit. However, if it
2297 finds options settings right at the start of the pattern, it extracts them into
2298 the global options, as before. Thus, they show up in the info data.
2299
2300 8. Added support for the \Q...\E escape sequence. Characters in between are
2301 treated as literals. This is slightly different from Perl in that $ and @ are
2302 also handled as literals inside the quotes. In Perl, they will cause variable
2303 interpolation. Note the following examples:
2304
2305 Pattern PCRE matches Perl matches
2306
2307 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2308 \Qabc\$xyz\E abc\$xyz abc\$xyz
2309 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2310
2311 For compatibility with Perl, \Q...\E sequences are recognized inside character
2312 classes as well as outside them.
2313
2314 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2315 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2316 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2317 signed/unsigned warnings.
2318
2319 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2320 option for pcretest, so I've replaced it by a simple function that does just
2321 that job.
2322
2323 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2324 "pcregrep -".
2325
2326 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2327 Java package. This provides some syntactic sugar for simple cases of what my
2328 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2329 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2330 item, you can use this simplified notation. Note that only makes sense with
2331 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2332 greediness, whatever the setting of the PCRE_UNGREEDY option.
2333
2334 13. A change of greediness default within a pattern was not taking effect at
2335 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2336 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2337 was abstracted outside.
2338
2339 14. PCRE now supports the \G assertion. It is true when the current matching
2340 position is at the start point of the match. This differs from \A when the
2341 starting offset is non-zero. Used with the /g option of pcretest (or similar
2342 code), it works in the same way as it does for Perl's /g option. If all
2343 alternatives of a regex begin with \G, the expression is anchored to the start
2344 match position, and the "anchored" flag is set in the compiled expression.
2345
2346 15. Some bugs concerning the handling of certain option changes within patterns
2347 have been fixed. These applied to options other than (?ims). For example,
2348 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2349 been the other way round. Some of this was related to change 7 above.
2350
2351 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2352 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2353 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2354 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2355
2356 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2357 mode. Unlike ".", it always matches newline, whatever the setting of
2358 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2359 assertions. Perl allows it, but it doesn't (in general) work because it can't
2360 calculate the length of the lookbehind. At least, that's the case for Perl
2361 5.8.0 - I've been told they are going to document that it doesn't work in
2362 future.
2363
2364 18. Added an error diagnosis for escapes that PCRE does not support: these are
2365 \L, \l, \N, \P, \p, \U, \u, and \X.
2366
2367 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2368 reading past the end of the pattern in cases such as /[abcd/.
2369
2370 20. PCRE was getting more memory than necessary for patterns with classes that
2371 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2372
2373 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2374 compiling PCRE for use with Virtual Pascal.
2375
2376 22. Small fix to the Makefile to make it work properly if the build is done
2377 outside the source tree.
2378
2379 23. Added a new extension: a condition to go with recursion. If a conditional
2380 subpattern starts with (?(R) the "true" branch is used if recursion has
2381 happened, whereas the "false" branch is used only at the top level.
2382
2383 24. When there was a very long string of literal characters (over 255 bytes
2384 without UTF support, over 250 bytes with UTF support), the computation of how
2385 much memory was required could be incorrect, leading to segfaults or other
2386 strange effects.
2387
2388 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2389 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2390 there was a subsequent back reference to those brackets. This meant that, for
2391 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2392 possible to check for precisely this case. All we can do is abandon the
2393 optimization if .* occurs inside capturing brackets when there are any back
2394 references whatsoever. (See below for a better fix that came later.)
2395
2396 26. The handling of the optimization for finding the first character of a
2397 non-anchored pattern, and for finding a character that is required later in the
2398 match were failing in some cases. This didn't break the matching; it just
2399 failed to optimize when it could. The way this is done has been re-implemented.
2400
2401 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2402
2403 28. Added a new feature that provides some of the functionality that Perl
2404 provides with (?{...}). The facility is termed a "callout". The way it is done
2405 in PCRE is for the caller to provide an optional function, by setting
2406 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2407 global variable. By default it is unset, which disables all calling out. To get
2408 the function called, the regex must include (?C) at appropriate points. This
2409 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2410 This provides a means of identifying different callout points. When PCRE
2411 reaches such a point in the regex, if pcre_callout has been set, the external
2412 function is called. It is provided with data in a structure called
2413 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2414 matching continues; if it returns a non-zero value, the match at the current
2415 point fails. However, backtracking will occur if possible. [This was changed
2416 later and other features added - see item 49 below.]
2417
2418 29. pcretest is upgraded to test the callout functionality. It provides a
2419 callout function that displays information. By default, it shows the start of
2420 the match and the current position in the text. There are some new data escapes
2421 to vary what happens:
2422
2423 \C+ in addition, show current contents of captured substrings
2424 \C- do not supply a callout function
2425 \C!n return 1 when callout number n is reached
2426 \C!n!m return 1 when callout number n is reached for the mth time
2427
2428 30. If pcregrep was called with the -l option and just a single file name, it
2429 output "<stdin>" if a match was found, instead of the file name.
2430
2431 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2432 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2433 pcre_exec(). This saves a malloc/free per call. The default value of
2434 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2435 when configuring.
2436
2437 32. The default maximum size of a compiled pattern is 64K. There have been a
2438 few cases of people hitting this limit. The code now uses macros to handle the
2439 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2440 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2441 configuring. Tests 2 and 5 work only with 2-byte links because they output
2442 debugging information about compiled patterns.
2443
2444 33. Internal code re-arrangements:
2445
2446 (a) Moved the debugging function for printing out a compiled regex into
2447 its own source file (printint.c) and used #include to pull it into
2448 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2449 separate copies.
2450
2451 (b) Defined the list of op-code names for debugging as a macro in
2452 internal.h so that it is next to the definition of the opcodes.
2453
2454 (c) Defined a table of op-code lengths for simpler skipping along compiled
2455 code. This is again a macro in internal.h so that it is next to the
2456 definition of the opcodes.
2457
2458 34. Added support for recursive calls to individual subpatterns, along the
2459 lines of Robin Houston's patch (but implemented somewhat differently).
2460
2461 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2462 allow it to read and process whole directories in Win32. This code was
2463 contributed by Lionel Fourquaux; it has not been tested by me.
2464
2465 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2466 used to name a group. Names consist of alphanumerics and underscores, and must
2467 be unique. Back references use the syntax (?P=name) and recursive calls use
2468 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2469 numbers. The function pcre_fullinfo() can be used after compilation to extract
2470 a name/number map. There are three relevant calls:
2471
2472 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2473 PCRE_INFO_NAMECOUNT yields the number of entries
2474 PCRE_INFO_NAMETABLE yields a pointer to the map.
2475
2476 The map is a vector of fixed-size entries. The size of each entry depends on
2477 the length of the longest name used. The first two bytes of each entry are the
2478 group number, most significant byte first. There follows the corresponding
2479 name, zero terminated. The names are in alphabetical order.
2480
2481 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2482 case instead of 255. Making it the same both with and without UTF-8 support
2483 means that the same test output works with both.
2484
2485 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2486 calling malloc() with a zero argument.
2487
2488 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2489 optimization. I've improved things by keeping a bitmap of backreferences with
2490 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2491 fact referenced, the optimization can be applied. It is unlikely that a
2492 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2493 the match to follow \n) will appear inside brackets with a number greater than
2494 31, but if it does, any back reference > 31 suppresses the optimization.
2495
2496 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2497 of disabling numbered capturing parentheses. Any opening parenthesis that is
2498 not followed by ? behaves as if it were followed by ?: but named parentheses
2499 can still be used for capturing (and they will acquire numbers in the usual
2500 way).
2501
2502 41. Redesigned the return codes from the match() function into yes/no/error so
2503 that errors can be passed back from deep inside the nested calls. A malloc
2504 failure while inside a recursive subpattern call now causes the
2505 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2506
2507 42. It is now possible to set a limit on the number of times the match()
2508 function is called in a call to pcre_exec(). This facility makes it possible to
2509 limit the amount of recursion and backtracking, though not in a directly
2510 obvious way, because the match() function is used in a number of different
2511 circumstances. The count starts from zero for each position in the subject
2512 string (for non-anchored patterns). The default limit is, for compatibility, a
2513 large number, namely 10 000 000. You can change this in two ways:
2514
2515 (a) When configuring PCRE before making, you can use --with-match-limit=n
2516 to set a default value for the compiled library.
2517
2518 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2519 a different value is set. See 45 below.
2520
2521 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2522
2523 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2524 of things that can be changed at compile time. The first argument specifies
2525 what is wanted and the second points to where the information is to be placed.
2526 The current list of available information is:
2527
2528 PCRE_CONFIG_UTF8
2529
2530 The output is an integer that is set to one if UTF-8 support is available;
2531 otherwise it is set to zero.
2532
2533 PCRE_CONFIG_NEWLINE
2534
2535 The output is an integer that it set to the value of the code that is used for
2536 newline. It is either LF (10) or CR (13).
2537
2538 PCRE_CONFIG_LINK_SIZE
2539
2540 The output is an integer that contains the number of bytes used for internal
2541 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2542
2543 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2544
2545 The output is an integer that contains the threshold above which the POSIX
2546 interface uses malloc() for output vectors. See item 31 above.
2547
2548 PCRE_CONFIG_MATCH_LIMIT
2549
2550 The output is an unsigned integer that contains the default limit of the number
2551 of match() calls in a pcre_exec() execution. See 42 above.
2552
2553 44. pcretest has been upgraded by the addition of the -C option. This causes it
2554 to extract all the available output from the new pcre_config() function, and to
2555 output it. The program then exits immediately.
2556
2557 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2558 order to support additional features. One way would have been to define
2559 pcre_exec2() (for example) with extra arguments, but this would not have been
2560 extensible, and would also have required all calls to the original function to
2561 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2562 is used for passing in "extra" data from pcre_study().
2563
2564 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2565 contains the following fields:
2566
2567 flags a bitmap indicating which of the following fields are set
2568 study_data opaque data from pcre_study()
2569 match_limit a way of specifying a limit on match() calls for a specific
2570 call to pcre_exec()
2571 callout_data data for callouts (see 49 below)
2572
2573 The flag bits are also defined in pcre.h, and are
2574
2575 PCRE_EXTRA_STUDY_DATA
2576 PCRE_EXTRA_MATCH_LIMIT
2577 PCRE_EXTRA_CALLOUT_DATA
2578
2579 The pcre_study() function now returns one of these new pcre_extra blocks, with
2580 the actual study data pointed to by the study_data field, and the
2581 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2582 before. That is, this change is entirely upwards-compatible and requires no
2583 change to existing code.
2584
2585 If you want to pass in additional data to pcre_exec(), you can either place it
2586 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2587 block.
2588
2589 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2590 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2591 times with different match limits, until it finds the minimum value needed for
2592 pcre_exec() to complete. The value is then output. This can be instructive; for
2593 most simple matches the number is quite small, but for pathological cases it
2594 gets very large very quickly.
2595
2596 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2597 returns the size of the data block pointed to by the study_data field in a
2598 pcre_extra block, that is, the value that was passed as the argument to
2599 pcre_malloc() when PCRE was getting memory in which to place the information
2600 created by pcre_study(). The fourth argument should point to a size_t variable.
2601 pcretest has been extended so that this information is shown after a successful
2602 pcre_study() call when information about the compiled regex is being displayed.
2603
2604 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2605 because what follows is always an absolute path. (Later: it turns out that this
2606 is more than cosmetic for MinGW, because it doesn't like empty path
2607 components.)
2608
2609 49. Some changes have been made to the callout feature (see 28 above):
2610
2611 (i) A callout function now has three choices for what it returns:
2612
2613 0 => success, carry on matching
2614 > 0 => failure at this point, but backtrack if possible
2615 < 0 => serious error, return this value from pcre_exec()
2616
2617 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2618 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2619 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2620 use by callout functions. It will never be used by PCRE itself.
2621
2622 (ii) The pcre_extra structure (see 45 above) has a void * field called
2623 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2624 pcre_callout_block structure has a field of the same name. The contents of
2625 the field passed in the pcre_extra structure are passed to the callout
2626 function in the corresponding field in the callout block. This makes it
2627 easier to use the same callout-containing regex from multiple threads. For
2628 testing, the pcretest program has a new data escape
2629
2630 \C*n pass the number n (may be negative) as callout_data
2631
2632 If the callout function in pcretest receives a non-zero value as
2633 callout_data, it returns that value.
2634
2635 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2636 there were some redundant $(CFLAGS) in commands that are now specified as
2637 $(LINK), which already includes $(CFLAGS).
2638
2639 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2640 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2641 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2642 one-byte characters throughout. Note that case-insensitive matching applies
2643 only to characters whose values are less than 256. PCRE doesn't support the
2644 notion of cases for higher-valued characters.
2645
2646 (i) A character class whose characters are all within 0-255 is handled as
2647 a bit map, and the map is inverted for negative classes. Previously, a
2648 character > 255 always failed to match such a class; however it should
2649 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2650
2651 (ii) A negated character class with a single character < 255 is coded as
2652 "not this character" (OP_NOT). This wasn't working properly when the test
2653 character was multibyte, either singly or repeated.
2654
2655 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2656 mode, for example: \x{100}{2,3}.
2657
2658 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2659 singly or repeated) now correctly test multibyte characters. However,
2660 PCRE doesn't recognize any characters with values greater than 255 as
2661 digits, spaces, or word characters. Such characters always match \D, \S,
2662 and \W, and never match \d, \s, or \w.
2663
2664 (v) Classes may now contain characters and character ranges with values
2665 greater than 255. For example: [ab\x{100}-\x{400}].
2666
2667 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2668 PCRE in UTF-8 mode.
2669
2670 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2671 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2672 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2673 value.)
2674
2675 53. The single man page has become too large. I have therefore split it up into
2676 a number of separate man pages. These also give rise to individual HTML pages;
2677 these are now put in a separate directory, and there is an index.html page that
2678 lists them all. Some hyperlinking between the pages has been installed.
2679
2680 54. Added convenience functions for handling named capturing parentheses.
2681
2682 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2683 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2684 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2685 are faulted.
2686
2687 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2688 calling configure. These values are used when compiling the dftables.c program
2689 which is run to generate the source of the default character tables. They
2690 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2691 you will need to set these values.
2692
2693 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2694
2695
2696 Version 3.9 02-Jan-02
2697 ---------------------
2698
2699 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2700
2701 2. If --disable-static was given, the building process failed when trying to
2702 build pcretest and pcregrep. (For some reason it was using libtool to compile
2703 them, which is not right, as they aren't part of the library.)
2704
2705
2706 Version 3.8 18-Dec-01
2707 ---------------------
2708
2709 1. The experimental UTF-8 code was completely screwed up. It was packing the
2710 bytes in the wrong order. How dumb can you get?
2711
2712
2713 Version 3.7 29-Oct-01
2714 ---------------------
2715
2716 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2717 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2718 this didn't happen under Solaris 8, where I normally test things.
2719
2720 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2721 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2722 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2723 this fix an hour or so after the initial 3.7 release.)
2724
2725
2726 Version 3.6 23-Oct-01
2727 ---------------------
2728
2729 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2730 offsets passed as NULL with zero offset count.
2731
2732 2. The config.guess and config.sub files had not been updated when I moved to
2733 the latest autoconf.
2734
2735
2736 Version 3.5 15-Aug-01
2737 ---------------------
2738
2739 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2740 had been forgotten.
2741
2742 2. By using declared but undefined structures, we can avoid using "void"
2743 definitions in pcre.h while keeping the internal definitions of the structures
2744 private.
2745
2746 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2747 user point of view, this means that both static and shared libraries are built
2748 by default, but this can be individually controlled. More of the work of
2749 handling this static/shared cases is now inside libtool instead of PCRE's make
2750 file.
2751
2752 4. The pcretest utility is now installed along with pcregrep because it is
2753 useful for users (to test regexs) and by doing this, it automatically gets
2754 relinked by libtool. The documentation has been turned into a man page, so
2755 there are now .1, .txt, and .html versions in /doc.
2756
2757 5. Upgrades to pcregrep:
2758 (i) Added long-form option names like gnu grep.
2759 (ii) Added --help to list all options with an explanatory phrase.
2760 (iii) Added -r, --recursive to recurse into sub-directories.
2761 (iv) Added -f, --file to read patterns from a file.
2762
2763 6. pcre_exec() was referring to its "code" argument before testing that
2764 argument for NULL (and giving an error if it was NULL).
2765
2766 7. Upgraded Makefile.in to allow for compiling in a different directory from
2767 the source directory.
2768
2769 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2770 options bits, the pointer it was passed was to an int instead of to an unsigned
2771 long int. This mattered only on 64-bit systems.
2772
2773 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2774 generated) instead of pcre.in, which it its source. Also made the same change
2775 in several of the .c files.
2776
2777 10. A new release of gcc defines printf() as a macro, which broke pcretest
2778 because it had an ifdef in the middle of a string argument for printf(). Fixed
2779 by using separate calls to printf().
2780
2781 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2782 script, to force use of CR or LF instead of \n in the source. On non-Unix
2783 systems, the value can be set in config.h.
2784
2785 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2786 absolute limit. Changed the text of the error message to make this clear, and
2787 likewise updated the man page.
2788
2789 13. The limit of 99 on the number of capturing subpatterns has been removed.
2790 The new limit is 65535, which I hope will not be a "real" limit.
2791
2792
2793 Version 3.4 22-Aug-00
2794 ---------------------
2795
2796 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2797
2798 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2799
2800
2801 Version 3.3 01-Aug-00
2802 ---------------------
2803
2804 1. If an octal character was given, but the value was greater than \377, it
2805 was not getting masked to the least significant bits, as documented. This could
2806 lead to crashes in some systems.
2807
2808 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2809 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2810
2811 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2812 These just pass their arguments on to (pcre_free)(), but they are provided
2813 because some uses of PCRE bind it to non-C systems that can call its functions,
2814 but cannot call free() or pcre_free() directly.
2815
2816 4. Add "make test" as a synonym for "make check". Corrected some comments in
2817 the Makefile.
2818
2819 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2820 Makefile.
2821
2822 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2823 command called pgrep for grepping around the active processes.
2824
2825 7. Added the beginnings of support for UTF-8 character strings.
2826
2827 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2828 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2829 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2830 out for the ar command.)
2831
2832
2833 Version 3.2 12-May-00
2834 ---------------------
2835
2836 This is purely a bug fixing release.
2837
2838 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2839 of ZA. This was just one example of several cases that could provoke this bug,
2840 which was introduced by change 9 of version 2.00. The code for breaking
2841 infinite loops after an iteration that matches an empty string was't working
2842 correctly.
2843
2844 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2845 when matched against abbab (for example). After matching an empty string, it
2846 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2847 caused it to match further down the string than it should.
2848
2849 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2850 was there because it doesn't seem to be needed, and it causes trouble on some
2851 systems, as it is not a Standard C header. It has been removed.
2852
2853 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2854 were reported on the Macintosh. The changes were from
2855
2856 while ((c = *(++ptr)) != 0 && c != '\n');
2857 to
2858 while ((c = *(++ptr)) != 0 && c != '\n') ;
2859
2860 Totally extraordinary, but if that's what it takes...
2861
2862 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2863 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2864 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2865 assumes the way PCRE uses memmove() (always moving upwards).
2866
2867 6. PCRE is being used in one environment where strchr() is not available. There
2868 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2869 faster code anyway.
2870
2871
2872 Version 3.1 09-Feb-00
2873 ---------------------
2874
2875 The only change in this release is the fixing of some bugs in Makefile.in for
2876 the "install" target:
2877
2878 (1) It was failing to install pcreposix.h.
2879
2880 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2881
2882
2883 Version 3.0 01-Feb-00
2884 ---------------------
2885
2886 1. Add support for the /+ modifier to perltest (to output $` like it does in
2887 pcretest).
2888
2889 2. Add support for the /g modifier to perltest.
2890
2891 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2892 matches null strings.
2893
2894 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2895 pattern. Perl treats empty patterns specially - it reuses the most recent
2896 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2897 effect.
2898
2899 5. The POSIX interface was broken in that it was just handing over the POSIX
2900 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2901 required a bigger vector, with some working space on the end. This means that
2902 the POSIX wrapper now has to get and free some memory, and copy the results.
2903
2904 6. Added some simple autoconf support, placing the test data and the
2905 documentation in separate directories, re-organizing some of the
2906 information files, and making it build pcre-config (a GNU standard). Also added
2907 libtool support for building PCRE as a shared library, which is now the
2908 default.
2909
2910 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2911 09 are not valid octal constants. Single digits will be used for minor values
2912 less than 10.
2913
2914 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2915 existing programs that set these in the POSIX interface can use PCRE without
2916 modification.
2917
2918 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2919 return all that pcre_info() returns, plus additional data. The pcre_info()
2920 function is retained for compatibility, but is considered to be obsolete.
2921
2922 10. Added experimental recursion feature (?R) to handle one common case that
2923 Perl 5.6 will be able to do with (?p{...}).
2924
2925 11. Added support for POSIX character classes like [:alpha:], which Perl is
2926 adopting.
2927
2928
2929 Version 2.08 31-Aug-99
2930 ----------------------
2931
2932 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2933 trying to match at the startoffset position, but instead was moving forward to
2934 the next newline as if a previous match had failed.
2935
2936 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2937 and could get into a loop if a null string was matched other than at the start
2938 of the subject.
2939
2940 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2941 be distinguished at compile time, and for completeness also added PCRE_DATE.
2942
2943 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2944 in GnuWin32 environments.
2945
2946
2947 Version 2.07 29-Jul-99
2948 ----------------------
2949
2950 1. The documentation is now supplied in plain text form and HTML as well as in
2951 the form of man page sources.
2952
2953 2. C++ compilers don't like assigning (void *) values to other pointer types.
2954 In particular this affects malloc(). Although there is no problem in Standard
2955 C, I've put in casts to keep C++ compilers happy.
2956
2957 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2958 should be (const char *).
2959
2960 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2961 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2962 However, I haven't made this a standard facility. The documentation doesn't
2963 mention it, and the Makefile doesn't support it.
2964
2965 5. The Makefile now contains an "install" target, with editable destinations at
2966 the top of the file. The pcretest program is not installed.
2967
2968 6. pgrep -V now gives the PCRE version number and date.
2969
2970 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2971 causing the entire string to be ignored, instead of just the last character.
2972
2973 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2974 non-matching string, it can take a very, very long time, even for strings of
2975 quite modest length, because of the nested recursion. PCRE now does better in
2976 some of these cases. It does this by remembering the last required literal
2977 character in the pattern, and pre-searching the subject to ensure it is present
2978 before running the real match. In other words, it applies a heuristic to detect
2979 some types of certain failure quickly, and in the above example, if presented
2980 with a string that has no trailing " it gives "no match" very quickly.
2981
2982 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2983 other alternatives are tried instead.
2984
2985
2986 Version 2.06 09-Jun-99
2987 ----------------------
2988
2989 1. Change pcretest's output for amount of store used to show just the code
2990 space, because the remainder (the data block) varies in size between 32-bit and
2991 64-bit systems.
2992
2993 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2994 start matching at. This allows lookbehinds to work when searching for multiple
2995 occurrences in a string.
2996
2997 3. Added additional options to pcretest for testing multiple occurrences:
2998
2999 /+ outputs the rest of the string that follows a match
3000 /g loops for multiple occurrences, using the new startoffset argument
3001 /G loops for multiple occurrences by passing an incremented pointer
3002
3003 4. PCRE wasn't doing the "first character" optimization for patterns starting
3004 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3005 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3006 the letter 'x'. On long subject strings, this gives a significant speed-up.
3007
3008
3009 Version 2.05 21-Apr-99
3010 ----------------------
3011
3012 1. Changed the type of magic_number from int to long int so that it works
3013 properly on 16-bit systems.
3014
3015 2. Fixed a bug which caused patterns starting with .* not to work correctly
3016 when the subject string contained newline characters. PCRE was assuming
3017 anchoring for such patterns in all cases, which is not correct because .* will
3018 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3019 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3020 must be retried after every newline in the subject.
3021
3022
3023 Version 2.04 18-Feb-99
3024 ----------------------
3025
3026 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3027 computation of the store needed to hold the pattern was incorrect (too large).
3028 If such patterns were nested a few deep, this could multiply and become a real
3029 problem.
3030
3031 2. Added /M option to pcretest to show the memory requirement of a specific
3032 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3033
3034 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3035 compiled in such a way that the backtracking after subsequent failure was
3036 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3037 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3038
3039
3040 Version 2.03 02-Feb-99
3041 ----------------------
3042
3043 1. Fixed typo and small mistake in man page.
3044
3045 2. Added 4th condition (GPL supersedes if conflict) and created separate
3046 LICENCE file containing the conditions.
3047
3048 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3049 Perl, that is the internal \ allows the delimiter to be included in the
3050 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3051 the final delimiter, add \ to the end of the pattern (to test the error).
3052
3053 4. Added the convenience functions for extracting substrings after a successful
3054 match. Updated pcretest to make it able to test these functions.
3055
3056
3057 Version 2.02 14-Jan-99
3058 ----------------------
3059
3060 1. Initialized the working variables associated with each extraction so that
3061 their saving and restoring doesn't refer to uninitialized store.
3062
3063 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3064 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3065 fix the problem.
3066
3067 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3068 calls, and wasn't printing the correct value for compiling calls. Increased the
3069 default value of LOOPREPEAT, and the number of significant figures in the
3070 times.
3071
3072 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3073
3074 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3075 a building problem on Windows NT with a FAT file system.
3076
3077
3078 Version 2.01 21-Oct-98
3079 ----------------------
3080
3081 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3082 to character tables built by pcre_maketables() in the current locale. If NULL
3083 is passed, the default tables are used.
3084
3085
3086 Version 2.00 24-Sep-98
3087 ----------------------
3088
3089 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3090 it any more.
3091
3092 2. Allow quantification of (?>) groups, and make it work correctly.
3093
3094 3. The first character computation wasn't working for (?>) groups.
3095
3096 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3097 end of the subject) and add 5.005's \z, which really does match only at the
3098 very end of the subject.
3099
3100 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3101
3102 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3103 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3104 localized options. All options to pcre_study() were also removed.
3105
3106 7. Add other new features from 5.005:
3107
3108 $(?<= positive lookbehind
3109 $(?<! negative lookbehind
3110 (?imsx-imsx) added the unsetting capability
3111 such a setting is global if at outer level; local otherwise
3112 (?imsx-imsx:) non-capturing groups with option setting
3113 (?(cond)re|re) conditional pattern matching
3114
3115 A backreference to itself in a repeated group matches the previous
3116 captured string.
3117
3118 8. General tidying up of studying (both automatic and via "study")
3119 consequential on the addition of new assertions.
3120
3121 9. As in 5.005, unlimited repeated groups that could match an empty substring
3122 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3123 runtime if any iteration does actually match an empty substring.
3124
3125 10. Include the RunTest script in the distribution.
3126
3127 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3128 discrepancies, some of which were old and were also with respect to 5.004. They
3129 have now been fixed.
3130
3131
3132 Version 1.09 28-Apr-98
3133 ----------------------
3134
3135 1. A negated single character class followed by a quantifier with a minimum
3136 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3137 program crashes, or just wrong answers. This did not apply to negated classes
3138 containing more than one character, or to minima other than one.
3139
3140
3141 Version 1.08 27-Mar-98
3142 ----------------------
3143
3144 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3145
3146 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3147 latter must appear before anything that relies on it in the pattern.
3148
3149
3150 Version 1.07 16-Feb-98
3151 ----------------------
3152
3153 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3154 repeat of a potentially empty string).
3155
3156
3157 Version 1.06 23-Jan-98
3158 ----------------------
3159
3160 1. Added Markus Oberhumer's little patches for C++.
3161
3162 2. Literal strings longer than 255 characters were broken.
3163
3164
3165 Version 1.05 23-Dec-97
3166 ----------------------
3167
3168 1. Negated character classes containing more than one character were failing if
3169 PCRE_CASELESS was set at run time.
3170
3171
3172 Version 1.04 19-Dec-97
3173 ----------------------
3174
3175 1. Corrected the man page, where some "const" qualifiers had been omitted.
3176
3177 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3178 input syntax.
3179
3180 3. Fixed memory leak which occurred when a regex with back references was
3181 matched with an offsets vector that wasn't big enough. The temporary memory
3182 that is used in this case wasn't being freed if the match failed.
3183
3184 4. Tidied pcretest to ensure it frees memory that it gets.
3185
3186 5. Temporary memory was being obtained in the case where the passed offsets
3187 vector was exactly big enough.
3188
3189 6. Corrected definition of offsetof() from change 5 below.
3190
3191 7. I had screwed up change 6 below and broken the rules for the use of
3192 setjmp(). Now fixed.
3193
3194
3195 Version 1.03 18-Dec-97
3196 ----------------------
3197
3198 1. A erroneous regex with a missing opening parenthesis was correctly
3199 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3200 on some systems.
3201
3202 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3203 it was reported that one broken compiler failed on the former because "code" is
3204 also an independent variable.
3205
3206 3. The erroneous regex a[]b caused an array overrun reference.
3207
3208 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3209 fail on data ending with that character. (It was going on too far, and checking
3210 the next character, typically a binary zero.) This was specific to the
3211 optimized code for single-character negative classes.
3212
3213 5. Added a contributed patch from the TIN world which does the following:
3214
3215 + Add an undef for memmove, in case the the system defines a macro for it.
3216
3217 + Add a definition of offsetof(), in case there isn't one. (I don't know
3218 the reason behind this - offsetof() is part of the ANSI standard - but
3219 it does no harm).
3220
3221 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3222 most of the places where whitespace preceded '#'. I have given up and
3223 allowed the remaining 2 cases to be at the margin.
3224
3225 + Rename some variables in pcre to eliminate shadowing. This seems very
3226 pedantic, but does no harm, of course.
3227
3228 6. Moved the call to setjmp() into its own function, to get rid of warnings
3229 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3230
3231 7. Constructs such as \d{8,} were compiling into the equivalent of
3232 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3233 outcome, but in this particular case used more store than had been allocated,
3234 which caused the bug to be discovered because it threw up an internal error.
3235
3236 8. The debugging code in both pcre and pcretest for outputting the compiled
3237 form of a regex was going wrong in the case of back references followed by
3238 curly-bracketed repeats.
3239
3240
3241 Version 1.02 12-Dec-97
3242 ----------------------
3243
3244 1. Typos in pcre.3 and comments in the source fixed.
3245
3246 2. Applied a contributed patch to get rid of places where it used to remove
3247 'const' from variables, and fixed some signed/unsigned and uninitialized
3248 variable warnings.
3249
3250 3. Added the "runtest" target to Makefile.
3251
3252 4. Set default compiler flag to -O2 rather than just -O.
3253
3254
3255 Version 1.01 19-Nov-97
3256 ----------------------
3257
3258 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3259 like /([ab]*)*/, that is, for classes with more than one character in them.
3260
3261 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3262 as /((?>a*))*/ (a PCRE_EXTRA facility).
3263
3264
3265 Version 1.00 18-Nov-97
3266 ----------------------
3267
3268 1. Added compile-time macros to support systems such as SunOS4 which don't have
3269 memmove() or strerror() but have other things that can be used instead.
3270
3271 2. Arranged that "make clean" removes the executables.
3272
3273
3274 Version 0.99 27-Oct-97
3275 ----------------------
3276
3277 1. Fixed bug in code for optimizing classes with only one character. It was
3278 initializing a 32-byte map regardless, which could cause it to run off the end
3279 of the memory it had got.
3280
3281 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3282
3283
3284 Version 0.98 22-Oct-97
3285 ----------------------
3286
3287 1. Fixed bug in code for handling temporary memory usage when there are more
3288 back references than supplied space in the ovector. This could cause segfaults.
3289
3290
3291 Version 0.97 21-Oct-97
3292 ----------------------
3293
3294 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3295
3296 2. Optimized negated single characters not to use a bit map.
3297
3298 3. Brought error texts together as macro definitions; clarified some of them;
3299 fixed one that was wrong - it said "range out of order" when it meant "invalid
3300 escape sequence".
3301
3302 4. Changed some char * arguments to const char *.
3303
3304 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3305
3306 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3307 pcretest.
3308
3309
3310 Version 0.96 16-Oct-97
3311 ----------------------
3312
3313 1. Added a simple "pgrep" utility to the distribution.
3314
3315 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3316 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3317 where "ddd" means "one or more decimal digits".
3318
3319 3. Fixed serious bug. If a pattern had a back reference, but the call to
3320 pcre_exec() didn't supply a large enough ovector to record the related
3321 identifying subpattern, the match always failed. PCRE now remembers the number
3322 of the largest back reference, and gets some temporary memory in which to save
3323 the offsets during matching if necessary, in order to ensure that
3324 backreferences always work.
3325
3326 4. Increased the compatibility with Perl in a number of ways:
3327
3328 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3329 to request this handling. The option can be set at compile or exec time.
3330
3331 (b) $ matches before a terminating newline by default; an option
3332 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3333 mode). The option can be set at compile or exec time.
3334
3335 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3336 the same as Perl's. If the decimal number it represents is less than 10
3337 or there aren't that many previous left capturing parentheses, an octal
3338 escape is read. Inside a character class, it's always an octal escape,
3339 even if it is a single digit.
3340
3341 (d) An escaped but undefined alphabetic character is taken as a literal,
3342 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3343 escapes.
3344
3345 (e) {0} is now permitted. (The previous item is removed from the compiled
3346 pattern).
3347
3348 5. Changed all the names of code files so that the basic parts are no longer
3349 than 10 characters, and abolished the teeny "globals.c" file.
3350
3351 6. Changed the handling of character classes; they are now done with a 32-byte
3352 bit map always.
3353
3354 7. Added the -d and /D options to pcretest to make it possible to look at the
3355 internals of compilation without having to recompile pcre.
3356
3357
3358 Version 0.95 23-Sep-97
3359 ----------------------
3360
3361 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3362 \x20 at the start of a run of normal characters. These were being treated as
3363 real characters, instead of the source characters being re-checked.
3364
3365
3366 Version 0.94 18-Sep-97
3367 ----------------------
3368
3369 1. The functions are now thread-safe, with the caveat that the global variables
3370 containing pointers to malloc() and free() or alternative functions are the
3371 same for all threads.
3372
3373 2. Get pcre_study() to generate a bitmap of initial characters for non-
3374 anchored patterns when this is possible, and use it if passed to pcre_exec().
3375
3376
3377 Version 0.93 15-Sep-97
3378 ----------------------
3379
3380 1. /(b)|(:+)/ was computing an incorrect first character.
3381
3382 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3383 but not actually doing anything yet.
3384
3385 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3386 as Perl does (e.g. [-az] or [az-]).
3387
3388 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3389 all possible positions.
3390
3391 5. Split up into different modules to avoid including unneeded functions in a
3392 compiled binary. However, compile and exec are still in one module. The "study"
3393 function is split off.
3394
3395 6. The character tables are now in a separate module whose source is generated
3396 by an auxiliary program - but can then be edited by hand if required. There are
3397 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3398 toupper() in the code.
3399
3400 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3401 make them global. Abolish the function for setting them, as the caller can now
3402 set them directly.
3403
3404
3405 Version 0.92 11-Sep-97
3406 ----------------------
3407
3408 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3409 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3410
3411 2. Caseless matching was not working in character classes if the characters in
3412 the pattern were in upper case.
3413
3414 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3415
3416 4. Make PCRE_ANCHORED public and accept as a compile option.
3417
3418 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3419 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3420 pass them.
3421
3422 6. Give an error if bad option bits passed at compile or run time.
3423
3424 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3425 pcretest to cause it to pass that flag.
3426
3427 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3428 options, and the first character, if set.
3429
3430 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3431
3432
3433 Version 0.91 10-Sep-97
3434 ----------------------
3435
3436 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3437 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3438
3439 2. PCRE was looping on encountering an indefinitely repeated back reference to
3440 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3441 Perl does - treats the match as successful.
3442
3443 ****

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